Cross-docking with IFD-MD

Tutorial Created with Software Release: 2023-2
Topics: Hit Discovery, Small Molecule Drug Discovery, Structure Prediction & Target Enablement
Products Used: IFD-MD

Tutorial files

8.4 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 you will learn how to use IFD-MD to generate a predicted binding pose of a known active compound using a holo crystal structure solved with a different ligand as a starting point.

 

Tutorial Content

  1. Introduction

  1. Creating Projects and Importing Structures

  1. Cross-docking with IFD-MD

  1. Analyzing IFD-MD Output with the Crystal Pose

  1. Increase Confidence in IFD-MD Predicted Poses

  1. Conclusion and References

  1. Glossary of Terms

1. Introduction

Many structure-based drug design methods, including free energy perturbation (FEP+), require accurate, atomic-level detail structures of the target protein complexed with a member of the ligand series being explored to perform optimally; consequently, the domain of applicability of structure-based drug design (SBDD) is limited by the availability of high-resolution crystal structures. Even when the exact protein-ligand complex structure needed for design work is not available, accurate atomic-level detail structures are often available for similar systems. There is often a significant lag between the identification of a hit and the structural-enablement of drug discovery work around that hit. A new ligand subseries sometimes results in a changed binding mode. IFD-MD enables structure-based drug discovery before the relevant co-crystal structure is available.

This tutorial will walk you through one such scenario where we want to enable design work around a compound with confirmed binding to thrombin, our known active, without access to any structures of that compound, or close analogues of it. We will use a structure solved in 2003 of thrombin with a different ligand bound in the binding site and use IFD-MD to predict the structure of our known active. A structure of our known active complexed with thrombin was solved in 2004, so our last step will be to compare the predicted structure of the binding site of our known active with this structure and see how much the design work was capable of without the structure solved in 2004.      

Often the only differences in protein structure between the available structure and the one that is needed for SBDD are the movements of a couple of side chains or a small loop motion. However, These small changes can have a large impact on the accessible ligand binding modes. IFD-MD uses a combination of docking algorithms, water thermodynamics, empirical scoring functions, implicit solvent force field energies, and explicit solvent metadynamics trajectories to explore the motions of the target protein and simultaneously determine their relative energetics. This technology makes it possible to create accurate “first looks” at the protein-ligand interactions for novel active compounds before a crystal structure is solved, and even allows for accurate structures of protein-ligand binding when starting from homology models. IFD-MD integrates multiple Schrödinger applications into a single solution for predicting binding poses, operating as a series of coupled CPU and GPU jobs:

 

Figure 1-1. IFD-MD Subjobs. Each dashed vertical line is a checkpoint. The job can be restarted from any checkpoint.

Initial Pose Generation (Phase, Prime, Glide)

Initial pose generation is done by first pharmacophore docking your ligand of interest (target ligand) onto the ligand of an existing holostructure ligand (template ligand). This ligand-based docking initially ignores clashes to the receptor. Pharmacophore docking generates thousands of conformations of your target ligand. For each conformation, clashes with the receptor are resolved using Prime to produce potentially a unique receptor conformation for each ligand conformation. The new receptor conformations are then re-docked using Glide to create an initial induced fit ligand receptor complex. For more details, please consult the references mentioned at the end of the tutorial.

Molecular Dynamic Stages

Concerted ligand-receptor changes are accomplished through short time scale (500 ps) MD simulations. The short MD time-scale also includes the placement of explicit water molecules, necessary for inclusion in scoring with Glide WS. The stability of a pose is evaluated using multiple metadynamics (MtD) calculations. Poses with more consistent conformations score more favorably. Poses with greater conformational variability score less favorably. IFD-MD incorporates this MtD stability as a component of its scoring function. A final 100 ns MD simulation is performed on the top two IFD-MD structures if the structure appears unconverged after 500 ps.

Note: Drastic backbone changes, i.e. DFG-in to DFG-out, are beyond the scope of IFD-MD predictions.

2. 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 2-2. 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-md_cross-docking.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 2-3. Opening IFD-MD tutorial project file.

  1. Go to File > Open Project
  2. Select the ifd-md.prjzip  and click Open
    • The pregenerated project opens as a 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 can be saved under a new name. Once a project is created, changes are saved automatically.

If you are unfamiliar with the Maestro GUI, please see the Glossary of Terms for the distinction between includedthe entry is represented in the Workspace, the circle in the In column is blue. and 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. entries in the Entry Lista simplified view of the Project Table that allows you to perform basic operations such as selection and inclusion..

Figure 2-4. Save Project panel.

  1. Go to File > Save Project As
  2. Change the File name to IFD-MD_tutorial
  3. Click Save.
    • The project is named IFD-MD_tutorial.prj

 

3. Cross-docking with IFD-MD

In the following example, we will assume that we ran a high-throughput, or virtual, screen and confirmed a novel hit for Thrombin, 1SL3. It was not until 2004 that 1SL3 was crystallized and confirmed active. Let us go back in time to 2003 and try to determine a binding pose of this confirmed active using the 1NZQ PDB structure as a template.

Before proceeding with an IFD-MD job, you must prepare the protein and ligands. For more information regarding the effect of protein and ligand preparations and their protocols, please see Sastry, G.M.; Adzhigirey, M.; Day, T.; Annabhimoju, R.; Sherman, W., J. Comput. Aid. Mol. Des., 2013, 27(3), 221-234 (DOI: 10.1007/s10822-013-9644-8).

3.1     Prepare 1NZQ complex

Figure 3-1. Including 1NZQ entry.

  1. Includethe entry is represented in the Workspace, the circle in the In column is blue. 1NZQ from Entry Lista simplified view of the Project Table that allows you to perform basic operations such as selection and inclusion. in the Workspacethe 3D display area in the center of the main window, where molecular structures are displayed.
    • As an alternative you can also import PDB 1NZQ with File > Get PDB

Figure 3-2. Launching Protein Preparation Workflow panel.

  1. In the Favorite Toolbar, click Protein Preparation
    • The Protein Preparation Workflow panel opens

 

Note: The provided .prjzip file also includes prepared protein structures if you wish to skip to section 3.2.

Figure 3-3. Preparation workflow for 1NZQ.

  1. Activate INTERACTIVE workflow to prepare the structure step-by-step.
    • The protein source is indicated in 1.Specify Protein
    • To read more about the different workflows in Protein Preparation Workflow panel check the panel help

Figure 3-4. Preprocessing 1NZQ.

  1. In 2.Preprocess step, expand More Options.
  2. Check Delete waters beyond hets and set the distance to 0.00 Å to delete all existing water molecules
    • Another way to delete all water molecules is from Quick Select from the toolbar by selecting Waters. Then right click on any of the selected atoms to delete them from Multiple Atoms Selected menu.
  3. Give your job a name or use the default
  4. Click Preprocess
    • When the job is done the output preprocessed structure is automatically 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. and 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.

IFD-MD uses Grand Canonical Monte Carlo (GCMC)1 and WaterMap2 to place waters around the ligand. If specific data is available to suggest that the water position is conserved between the input structure and the structure with the new ligand, the water can be kept in place, but otherwise it should be deleted so that IFD-MD can position waters appropriately for the new ligand.

Figure 3-5. Checking 1NZQ structures.

  1. In 3.Diagnostic and Analyze click Check Structure
    • You can also switch to the Substructures tab

Figure 3-6. Deleting chains D and L.

  1. In Substructures, click Ctrl+click (Cmd+click on Mac) chains D and L
    • Chains D and L are selected in the Workspace
  2. Click Delete from Entry

Note: Deleting these chains that are far away from the ligand binding site will not affect the results, but will speed up the calculations.

  1. If Ligand SIN 55 remains, select it in the table and click Delete from Entry.
  2. Click Workflow to return back to the Preparation Workflow tab

Figure 3-7. Optimizing H-bond.

  1. In 4.Optimize H-bond Assignments expand Settings and check Minimize hydrogens of altered species
  2. Click Optimize

Figure 3-8. Minimizing 1NZQ receptor.

  1. In 5.Minimize and Delete Waters expand Settings and check Optimize hydrogens only
  2. Click Clean Up
    • This allows relaxation of the H-bond network
  3. Close the panel

3.2 Prepare ligand templates

Figure 3-9. Preparing the 1SL3 ligand.

  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. 1SL3_ligand from the Entry Lista simplified view of the Project Table that allows you to perform basic operations such as selection and inclusion.
    • This is an SDF structure file for 1SL3 ligand downloaded from rcsb.org
  2. Open the LigPrep panel from Tasks
  3. For Use structures from, choose Project Table (1 selected entry)
  4. Change the Job name to ligprep_1sl3
  5. Run the job with default settings.
    • Job takes ~1 min

Figure 3-10. LigPrep job output structure.

  1. Close the LigPrep panel
    • The LigPrep job returns one structure as seen in Figure 3-10

3.3     Running IFD-MD cross-docking job from the panel

Figure 3-11. Launching IFD-MD panel.

  1. From Tasks, search for ifd
    • Two items are displayed in the results
  2. Select IFD-MD
    • The IFD-MD panel opens

Figure 3-12. Preparing IFD-MD job.

  1. Includethe entry is represented in the Workspace, the circle in the In column is blue. 1NZQ - 5-removed waters complex 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 1SL3_ligand of the ligprep_1sl3_lig-out group
  2. Next to Target ligand, choose Project Table and click Load
  3. Next to Template Complex from Workspace click Load
  4. Ligand H:162 (248) should auto-populate. If it does not, Pick the ligand in the Workspacethe 3D display area in the center of the main window, where molecular structures are displayed.

Figure 3-13. Adjusting IFD-MD job settings.

  1. Click cog button in the panel
    • Job settings panel opens
  2. Under Job, change Name to IFD-MD_1NZQ_with_1SL3
  3. Choose a CPU subhost
  4. Indicate the number of processors for the CPU subhost to use (50 CPUs were used in this tutorial)
  5. Choose a GPU subhost
  6. Indicate the number of GPUs for the GPU subhost to use (8 GPUs were used in this tutorial)
    • The panel runs one GPU per subjob
  7. As this is a lengthy job, do not click Run. However, clicking Run will submit this job from your local computer.
    • As the job takes many hours, we will use pre-generated results in the IFD-MD_1NZQ_with_1SL3_2-out group from the Entry Lista simplified view of the Project Table that allows you to perform basic operations such as selection and inclusion.

4. Analyzing IFD-MD Output with the Crystal Pose

The final output of the IFD-MD run for the predicted binding pose of 1NZQ with the 1SL3 ligand should yield 5 output structures ordered by IFD-MD Score. The first two structures have water molecules as they have been further refined by MD; the rest of the output structures did not undergo MD refinement and only have partial scoring. The top pose should have the lowest IFD-MD Score.

Figure 4-1 and 4-2. White: native structure of 1SL3 ligand ( the right answer that we have blinded ourselves from in this tutorial)

Pink - Starting structure

Gold - IFD-MD top ranked structure.

  • Most of the residues will not have changed, but GLU192 will have
    • It will have moved to accommodate the larger ligand
  • This allows the larger ligand to dock so we can establish that we are getting the right answer for the right reasons  

The following figures are a side-by-side comparison of the crystallographically determined pose (left) and the IFD-MD pose for a thrombin ligand (right).  Side chains are shown only for key residues.

 

 

Figure 4-3. Comparison of crystallographically determined pose ( left) and IFD-MD pose for a thrombin ligand (right) showing the correct vector for the solubilizing group.  The protein is represented as a surface,  the ligand is in bond representation.

5. Increase Confidence in IFD-MD Predicted Poses

IFD-MD generates poses that are then used as input for FEP+. The IFD-MD combined score is used to prioritize structures to be used with FEP+.

FEP+ can provide accurate predictions of binding affinity, assuming the IFD-MD structure has a low RMSD from the crystal structure of the structure with the congeneric ligand.

As next steps, use FEP+ for top scoring pose prediction by following the BACE1 Inhibitor Design Using Free Energy Perturbation tutorial.

6. Conclusion and References

In this tutorial, we utilized the binding poses of other ligands as a template to help determine the binding pose of the ligand of interest. Overall, the four main stages for the IFD-MD cross-docking protocol:

  1. Prepare the template complex that will be used as a starting point
  2. Prepare the known active ligand that will be docked
  3. Score with IFD-MD
  4. Compare the result  

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.