Glide WS Evaluation of Hsp90 Ligands

Tutorial Created with Software Release: 2025-3
Topics: Hit Discovery, Hit-to-Lead & Lead Optimization, Small Molecule Drug Discovery, Virtual Screening
Products Used: Glide

Tutorial files

53 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 displayedthe 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, you will learn how to use Glide WS to screen a set of ligands against two conformationally distinct structures for Hsp90. You will prepare the structures, generate the Glide WS ensemble model, set up the docking and analyze the results of the calculation.

Tutorial Content

  1. Creating Projects and Importing Structures

  1. Preparing Structures for Glide WS

  1. Generating Glide WS Models

  1. Docking with Glide WS

  1. Analyzing Docking Results with the Glide WS Visualizer

  1. Conclusion and References

  1. Glossary of Terms

1. Creating Projects and Importing Structures

At the start of the session, change the file path to your chosen Working Directorythe location that 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 created, the project is automatically saved each time a change is made.

Structures can be imported from the PDB directly, or from your Working Directorythe location that files are saved using File > Import Structures, and are added to the Entry Lista 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 Entry Lista 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 1-1. 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/glidews_hsp90.zip
  4. After downloading the zip file, unzip the contents in your Working Directorythe location that files are saved for ease of access throughout the tutorial.

Figure 1-2. Save Project panel.

  1. Go to File > Save Project As
  2. Change the File name to HSP90_GlideWS, click Save
    • The project is now named HSP90_GlideWS.prj

2. Preparing Structures for Glide WS

Glide WS uses an advanced, hard-scoring function, and is sensitive to improper protein preparation. When preparing structures for Glide WS, keep in mind that you will need to run WaterMap calculations whose results are included in the scoring. As molecular dynamics simulations will be run on the system, it is important that all valences are properly filled and there are no missing side chains. Breaks in the protein backbone may be tolerated, but can strongly influence hydration patterns in the binding pocket if they are nearby. As the final preparation step, you should remove all crystallographic water molecules and align the receptor structures using Binding Site Alignment. See the Introduction to Structure Preparation and Visualization tutorial and the Best Practices for Protein Preparation for more details on this process.

In this tutorial, you will be working with two high-quality structures for which we recommend the default settings of the Protein Preparation Workflow and removing all crystallographic water at the end. Note that using multiple receptor structures is optional, Glide WS can be used for scoring ligands against a single receptor.

Figure 2-1. Download 2FWZ and 2BT0 structures from the PDB.

  1. Go to File > Get PDB
  2. Next to PDB ID, type 2FWZ, 2CDD
  3. Next to Chain name, type A
  4. Click Download
    • Two structures are added to the Entry Lista simplified view of the Project Table that allows you to perform basic operations such as selection and inclusion.
  5. In the Entry Lista simplified view of the Project Table that allows you to perform basic operations such as selection and inclusion, includethe entry is represented in the Workspace, the circle in the In column is blue 2FWZ.

Figure 2-2. Specify the Protein in the Protein Preparation Workflow.

  1. Open the Protein Preparation Workflow from the Favorites Toolbar.
  2. In the Preparation Workflow tab, confirm the INTERACTIVE button is off.
  3. In the Specify Protein section, choose use structures from: Workspace.

The interactive mode can be used for exploring manual options, or to run a single protein in a step-by-step manner. This is especially important when working with proteins with uncommon protonation patterns or hydrogen bond networks.

In this tutorial, you will use an automatic protein preparation protocol as manual adjustments are not needed for these protein structures.

Figure 2-3. Preprocess structure with the Protein Preparation Workflow.

Any missing or incomplete side chains need to be added in the preprocessing stage. Any parts of the protein reconstructed during this stage need to be sanity-checked before proceeding, especially if they are near the binding site.

In the 2FWZ structure, the side chain of GLU 16 is missing and will be filled in by default. There are no missing loops you would need to account for.

  1. Confirm Preprocess is toggled on and Fill in missing side chains is checked.

 

When applying this to your own system, you can enable the option to fill in missing loops from the More Options button. See the Protein Preparation Workflow documentation page for more details on how this reconstruction works and its limitations.

Before rebuilding loops in this manner, pay attention to their location. If they are far from the binding site, their absence will not affect the docking, so there is no need to model them. If they are near the binding site, you should take great care when deciding on whether to rebuild them, as they will be considered to be fully rigid by Glide and may thus artificially constrain the available ligand poses.

We recommend additional validation before using a structure with loops reconstructed in this manner before docking. It is best practice to at least ensure that the co-crystal ligand re-docks in its correct pose.

Figure 2-4. Optimize H-bond Assignments with the Protein Preparation Workflow.

  1. Confirm Optimize H-bond Assignment is toggled on.
  2. Click Settings.
  3. Confirm Minimize hydrogens of altered species is unchecked.

Note: If you experience issues when re-docking the cocrystal ligand, it may be helpful to re-prepare your system with this optional minimization step enabled.

After the H-bond network has been optimized, all crystallographic water molecules should be removed from the structure. Some of the crystallographic waters are where they are due to interactions with the ligand, which is not present during the WaterMap simulation. In such a case, starting with a dry complex allows the system setup step in WaterMap to fill the ligand void as needed to avoid artifacts in the hydration site sampling.

For the docking stage of the GlideWS workflow, water molecules in the pocket are not supported and automatically deleted by the backend. To avoid confusion, we therefore recommend deleting all water molecules manually prior to running the GlideWS model generation.

Figure 2-5. Perform a Restrained Minimization with the Protein Preparation Workflow and run the batch rob.

  1. Confirm Minimize and Delete Waters is toggled on.
  2. Click Settings
  3. Check Delete Waters Distant from Ligands
  4. Change the distance cutoff for deleting waters to 0 Å.
  5. Change the job name to 2FWZ_Prepared
  6. Click Run
    • This job will take a few minutes. Feel free to move to the next step in the tutorial in the meantime.

Figure 2-6. Preparing the 2CDD structure.

Prepare the 2CDD structure in the same way.

  1. In the Entry Lista simplified view of the Project Table that allows you to perform basic operations such as selection and inclusion, includethe entry is represented in the Workspace, the circle in the In column is blue 2CDD
  2. Repeat the preparation steps 7 - 16 on 2CDD and name the job 2CDD_Prepared.

Note: This structure is missing the backbone for LYS 224, but that residue is nowhere near the binding site so no detailed inspection of the rebuilt residue is needed.

What if my structure has a membrane?

If the binding site is not exposed to the membrane, it should be fine not to include the membrane in the structure due to the spatial boundaries of the docking grid. Otherwise, you would need to carefully model the nearby membrane as part of the system.

Figure 2-7. Aligning the prepared structures based on their binding sites.

Before running the Glide WS calculation, all receptor structures need to be aligned. We recommend using Binding Site Alignment, as this part of the receptor is the focus for the docking algorithm.

  1. Includethe entry is represented in the Workspace, the circle in the In column is blue both prepared structures.
  2. Go to Tasks > Browse > Protein Preparation and Refinement > Binding Site Alignment.
  3. For Use proteins from, choose Workspace.
  4. Change the Job name to align_binding_sites_hsp90 and click Run.
    • This job should finish in a minute.
    • You can find the results in the outputs folder of the tutorial zip archive as align_binding_sites_hsp90-align-final.maegz.

Figure 2-8. Applying the Binding Mode comparison preset to the aligned structures.

You can now inspect and compare the binding sites.

  1. Includethe entry is represented in the Workspace, the circle in the In column is blue the aligned structures.
  2. From the Presets menu, choose Binding mode Comparison.

The two receptor structures you have just prepared are quite different: In the 2FWZ structure, the residues between ILE 104 and ILE 110 are part of a longer alpha helix. In the 2CDD structure, they are part of a flexible loop. This flexibility would not be captured in a regular Glide docking, resulting in filtering out actives which favor a different binding mode than the receptor for which the grid was generated.

Figure 2-9. Conformational differences near the binding site for the 2FWZ (white) and 2CDD (green) structures.

3. Generating Glide WS Models

Once all protein-ligand complexes have been prepared, they are ready for use in generating a Glide WS model. We will define the complexes that will be used to build the ensemble Glide WS will use for screening as well as the corresponding WaterMap data. The output is a model archive analogous to a Glide grid, but containing multiple receptor conformations.

For more information on running WaterMap calculations, see the Identifying Binding Site Requirements and Lead Optimization with WaterMap tutorial.

Figure 3-1. Glide WS Model Generation in the Tasks menu.

  1. Includethe entry is represented in the Workspace, the circle in the In column is blue and 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 Entry List (and Project Table) and the row for the entry is highlighted. Project operations are performed on all selected entries the prepared and aligned proteins in your Entry Lista simplified view of the Project Table that allows you to perform basic operations such as selection and inclusion
  2. Go to Tasks > Glide > Glide WS Model Generation
    • The Glide WS Model Generation panel opens

Note: The Glide WS Model Generation Panel documentation page has additional detailed information for how to set up Glide WS models and you should read it in detail before applying this to your own systems.

For each of the complexes you are working with, you now have to either provide the corresponding WaterMap data or specify that a WaterMap should be created. In this tutorial, we have provided pre-generated WaterMaps for both complexes.

 

You also need to specify the binding affinity for the reference ligand, as it is used to estimate the offset between the docking scores for multiple receptors. If you have the associated binding data for the complexes you are working with, you can load that data in when you initially add the entries to the Glide WS Model Generation panel. If no binding data is available, you should only work with a single receptor at a time, and specify a dummy value of -12 kcal/mol (approximately 1 nM). Using a value of 0 for the binding affinity is not recommended.

Figure 3-2. Load 2FWZ WaterMap data into the Glide WS Model Generation panel.

  1. Click Load.
  2. Select 2FWZ_-_prepared in the table
  3. In the Binding Free Energy text box, type -10.6.
  4. Click the checkmark button to commit the binding free energy value.
  5. Select WaterMap: From File and click “...”
  6. Locate 2FWZ_-_prepared_wm.maegz and click Open.

Figure 3-3. Load 2CDD WaterMap data into the Glide WS Model Generation panel.

Repeat the same steps for the 2CDD structure:

  1. Select 2CDD_-_prepared in the table
  2. In the Binding Free Energy text box, type -8.34.
  3. Click the checkmark button to commit the binding free energy value.
  4. Select WaterMap: From File and click “...”
  5. Locate 2CDD_-_prepared_wm.maegz and click Open.

Figure 3-4. Set-up Glide WS Model Generation job.

Now you can submit the job as usual.

  1. Set Job name to glide_ws_model_hsp90
  2. This job can take several days depending on resources and job settings, so we will be using the pre-generated file glide_ws_model_hsp90_models.zip going forward.

 

Note: If any WaterMap models are set to “Create”, you will need to specify a Linux GPU host to run the calculations by clicking the Cog button and choosing Job settings. See the WaterMap System Requirements.

4. Docking with Glide WS

The Glide WS model describes an ensemble of receptors, unlike a Glide Pose Viewer file which only describes a single receptor. During screening, each ligand will be docked into all of the available receptors in the ensemble, with the best scoring pose written to the .epv file.

Figure 4-1. Docking in Glide WS.

  1. Go to Tasks > Applications > Glide > Glide WS Docking
    • The Glide WS - Docking panel opens

Figure 4-2. Import model file into the Glide WS - Docking panel.

  1. Next to File, click Browse and select glide_ws_model_hsp90_models.zip
  2. Click Open

Next, we need to specify the ligands to be docked. The 1000 unique ligands used in this tutorial are a mix of known actives and decoys from the DUD-E dataset for HSP90. You can recognize the actives by their CHEMBL IDs. We have prepared these ligands using LigPrep, resulting in 1972 compounds for docking due to the expansion of accessible tautomer and protonation states. See the Introduction to Structure Preparation and Visualization tutorial to learn how this works.

Figure 4-3. Specify the ligands to be docked.

  1. Next to Use ligands from, choose File
  2. Next to File name, click Browse, choose hsp90_dude_epikx_sample_1k.maegz and click Open.
  3. Change Job name to Glide_WS_dock_HSP90
  4. This job can take several days depending on resources and job settings, so we will be using the pre-generated file Glide_WS_dock_HSP90_model1_epv.maegz going forward.

 

Note: If you run the job yourself, the pre-generated files may show differences compared to your results due to differences in software version, hardware configuration, or stochastic effects.

5. Analyzing Docking Results with the Glide WS Visualizer

Figure 5-1. Import pre-generated Glide WS docking results.

  1. Go to File > Import Structures
  2. Select Glide_WS_dock_HSP90_model1_epv.maegzand click Open
    • A new group is added to the Entry List.

Figure 5-2. Styling ligands in selected entries.

For the visual inspection of the poses, feel free to adjust visualization settings to your preference or skip this and continue from step 5.

 

  1. 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 Entry List (and Project Table) and the row for the entry is highlighted. Project operations are performed on all selected entries the Glide_WS_dock_HSP90_model1_epv.mae group.
  2. Use the Style selected entries icon to show the ligands as thin tubes.

You can see which receptor gave the best results for a given ligand by adding the best_recep property to the Entry List table.

  1. Click the three dots in the Entry List heading and choose Show Properties.
  2. Click Choose.
  3. Search for and click the best recep property.
  4. Click OK.
    • The best recep column is added to the Entry List.

Note: You may need to drag the edge of the table to make the column visible.

The Glide WS output is split into two groups, depending on whether MM-GBSA was run for that compound or not. To reduce the computational cost of the screening, calculating the MM-GBSA penalty is skipped for compounds for which the raw Glide WS scores are worse than a set threshold.

 

Here, MM-GBSA penalties were only calculated for about 20% of the compounds (those in the Glide WS Output group), the less promising compounds where this was omitted are in the Skipped MMGBSA group. A quick scroll through the latter group shows that only a few known binders ended up there. While looking at these cases can be helpful later for understanding the model behavior in detail, e.g. for troubleshooting purposes. However, your first focus should go to the ligands in the Glide WS Output group, for which all components of the Glide WS score were calculated.

Figure 5-3. Open the Glide WS Visualizer.

  1. Click the funnel workflow action menu next to the group header for the Glide WS Output.
    • The Glide WS Visualizer opens

Now, you can investigate the results. In the Entry List, the poses are sorted by Total Glide WS.

The Glide WS Visualizer shows both the total score for each pose and its components. There is an associated Workspace visualization for all properties that have a blue eye icon next to them.

The information symbols in these rows offer more insight on which atoms contribute to this term.

For a full overview of the Glide WS terms, see the most recent publication (you can find additional references in the Further Reading section below).

Figure 5-4. Overlay WaterMap sites in the Workspace.

You can investigate how each ligand occupies the WaterMap sites. For more details on interpreting WaterMap results, see the Target Analysis with SiteMap and WaterMap tutorial.

 

  1. Toggle on WaterMap sites
    • The WaterMap sites are now visible in the Workspace as spheres colored by their energy penalty.
  2. Hover over or click the information symbol next to the Water Displacement Reward in the visualizer to see which atoms contribute to this term.

Figure 5-5. Inspect Glide WS terms in the Workspace.

You can also analyze individual hydrogen bonds in more detail.

  1. Expand the Hydrogen Bond Score group
  2. Click the information symbol for the corresponding row.
    • The Workspace is zoomed to display the hydrogen bond donor and acceptor as well as the distance between them.

Figure 5-6. Toggle through Glide WS - Docking outputs.

  1. Use your right arrow key or click Next to flip through the ligands.
    • The visualizer automatically shows the best receptor for each ligand.

By stepping through the ligands and investigating the Glide WS terms, you can identify patterns in the beneficial and detrimental contributions. For example, interactions with the bridging water at ASP 93 are rewarded by the 'charged ligand motif' term. This interaction is the main discriminator between the known actives (identifiable by their CHEMBL IDs) and decoys.

 

When reviewing compounds in this project with poor docking scores, observe that many of these are being penalized for displacing WaterMap predicted hydration sites without forming compensatory hydrogen bonds, effectively desolvating polar groups on the receptor. These types of violations are not seen in crystallized complexes and are an effective way to discriminate actives from inactive compounds. This important type of analysis is only possible in Glide WS.

6. Conclusion and References

In this tutorial, we prepared two protein structures, used those structures to generate a Glide WS model, ran a screen with Glide WS, and finally analyzed the docking outputs with the Glide WS Visualizer. The receptor conformations chosen for this study are quite different from one another. By including multiple structures in the ensemble, we are able to capture different binding modes in a single screen. HSP90 contains a complex network of waters in the binding site. Glide WS detects these crucial waters, recognizing and rewarding or penalizing interactions that displace waters from the binding site.

For further reading:

 

 

7. Glossary of Terms

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

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

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 Entry List (and Project Table) and the row for the entry is highlighted. Project operations are performed on all selected entries

Working Directory - the location that files are saved

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