2022 DOCK tutorial 2 with PDBID 4ZUD
- 1 Introduction
- 2 Protein and Ligand Preparation
- 3 Surface Spheres Generation
- 4 Grid Box
- 5 DOCKING
- 6 Virtual Screening
DOCK is a molecular modeling program capable of sampling lower-energy ligand conformations with respect to a binding surface on a given protein. DOCK utilizes and manipulates the geometry of the ligand to find the conformation that yields that most favorable interaction with the respective binding site. With this tool, millions of molecules can be rapidly screened against a target protein for the purposes of identifying new drug molecules that are physiologically relevant.
For more information on DOCK and it's uses, please refer to their online manual: DOCK6 Manual
Before beginning, create the following directories in your space so that all necessary files are organized and can be access quickly:
mkdir 001.structure 002.surface_spheres 003.gridbox 004.dock 005.virtual_screen 006.cartesian_min 007.rescore
You don't have to name your directories the same as they are named here, but be cautious since the files that will be used for this tutorial utilize this naming scheme. They will need to be changed in each file that refers to them if you don't use this naming scheme!
Be sure to have Chimera installed on your system as it will be our primary visualization and system-editing program.
Protein and Ligand Preparation
Download the 4ZUD PDB file from the RCSB PDB website and open the file in Chimera.
Select -> Open -> (pathway to pdb file on your local machine) -> Open
You will notice a few side chain residues are explicitly displayed; those are the ones that directly engage with the ligand. The structure also has some missing regions denoted by the dashed-lines. These regions do not have to be modeled to use the system for docking since the majority of the protein remains restrained during the process (except for the residues of the active site, to a certain extent). You can play around with Chimera and visualize the protein from different angles to get a complete look at the protein to ensure there are no glaring errors in the structure that could have somehow arose from the downloading and opening process (Doesn't usually happen, but it's always good to be sure before moving on!)
Many structures deposited in the PDB lack hydrogens due to the difficulty in resolving their electron densities from cryo-EM or X-ray crystallography. The structures also lack formal charges since that information is not captured with out current experimental structure-determining techniques. Both charges and hydrogens are crucial for accurately studying any chemical system, and so they both must be added manually to 4ZUD in order to prime the system for docking.
First we want to delete the ligand from the file since we want to save the protein separately
Select -> Residue -> OLM Actions -> Atoms/Bonds -> Delete
4ZUD Structure Caveat
In the 4ZUD paper it mentions that ILE53 was mutated to an Alanine, but when you load the PDB into Chimera, it is recognized as an isoleucine but with an alanine side-chain.
There is no immediately obvious explanation for this, but since it is being recognized as an isoleucine, we're just going to edit the side chain to be the correct one. ILE53 can be selected by using the following command on the Chimera Command Line interface which could be accessed under Tools:
Tools -> General Controls -> Command Line
Make sure that you're selecting model 1 or whatever your current model is ranked at under Tools -> General Control -> Model Panel
Once residue 53 is selected, you can change the side-chain atoms by selecting a new rotamer type:
Tools -> Structure Editing -> Rotamers
In the Choose Rotamer Parameters window, select the ILE rotamer type and press apply.
In the ILE 53.A Side-Chain Rotamers window, choose the highest probability rotamers and press apply.
You should now see your alanine side chain change into an isoleucine one. It will remove the hydrogens so you have to add those back in, just for this residue.
To select the entire protein:
Select -> Chain -> A
To add the hydrogens to the protein:
Tools -> Structure Editing -> AddH
All residues in the protein should now have all of the hydrogens that were missing. Make sure to look at the output log of this command just in-case any errors arise, although there should be none if the instructions were followed thus far.
There is a similar Chimera command to add charges to your protein selection:
Tools -> Structure Editing -> Add Charge
After using this command you should receive an error stating that Correct charges are unknown for 3 non-standard atom names in otherwise standard residues. If you look at those atoms in the reply log, they're hydrogens belonging to ILE53. If you take a look at the paper that accompanies the 4ZUD structure, those hydrogens were replaced by tritium for crystallization purposes and Chimera does not recognize them as standard atoms and doesn't have predefined partial charges for them. Since Chimera doesn't recognize them, it will not apply charges to them. Additionally, since they are just hydrogens (meaning that their charge contributions to the system is often times very minimal) and the residue is not near the ligand active site, we do not have to do anything further in terms of adding charges.
The Add Charge command predicts the protein net charge to be -2.913. Make sure that this value is in agreement with what the PDB website or the corresponding paper says about the charge, if that information is provided.
Once you have completed all of the aforementioned steps, you have to save the protein as a mol2 file.
File -> Save Mol2
For the purposes of this tutorial the file will be called 4ZUD_protein_hydrogens.mol2
You will want to save a version of this file without hydrogens. To do that, you can select and delete all of the hydrogens like so:
Select -> Chemistry -> element -> H Actions -> Atoms/Bonds -> Delete
You can then resave the file. 4ZUD_protein_without_hydrogens.mol2
Now we want to focus on preparing the ligand. We can reopen the unmodified 4ZUD PDB in Chimera and delete the protein from the file. You can do this by doing an inverse selection for the protein. First select the ligand:
Select -> Residue -> OLM
On your keyboard, press Shift + Right-Arrow keys simultaneously to invert the selection to the parts that belong to the protein. You can then delete your selection, and you should be left with just the ligand:
Actions -> Atoms/Bonds -> Delete
To add hydrogens to the ligand, you can follow the same procedure that you used to add them to the protein § Adding Hydrogens.
You can follow the same procedures to add the charges to the ligand as well § Adding Charges.
The Add Charge command predicts a net -1 charge for the ligand. This makes sense since there is a carboxylate group in the molecules. Again, make sure to corroborate this with what the paper states and/or the PDB website states about the charge.
You can then save the file the same way you did for the protein mol2. For the purposes of this tutorial, the file will be named 4ZUD_ligand_hydrogens.mol2
Before going any further, make sure to copy the mol2 files that were generated thus far from your local computer to Seawulf. Place these files in the 001.structure directory, or whichever one is the equivalent directory in your records.
Surface Spheres Generation
UNDER CONSTRUCTION -- FILE NAMES, DIRECTORIES ARE FROM AS. WILL CHANGE THEM TO JA CONVENTIONS.
Next, we'll need to generate the grid we'll use in subsequent steps. Let's start by creating our grid.in file:
Into this file copy-paste the following:
compute_grids yes grid_spacing 0.4 output_molecule no contact_score no energy_score yes energy_cutoff_distance 9999 atom_model a attractive_exponent 6 repulsive_exponent 9 distance_dielectric yes dielectric_factor 4 bump_filter yes bump_overlap 0.75 receptor_file ../001_structure/4zud_receptor_wcharges_wH.mol2 box_file 4zud.box.pdb vdw_definition_file /gpfs/projects/AMS536/zzz.programs/dock6.9_release/parameters/vdw_AMBER_parm99.defn score_grid_prefix grid
Now that there is a grid.in file, the grid can be produced by running the following:
grid -i grid.in -o gridinfo.out
Upon successful completion of the grid generation, there should be three new files in your working directory: grid.bmp, grid.nrg, and gridinfo.out.
MY TUTORIAL LACKED THIS STEP -- DO WE NEED THIS?
The main goal of this section is to conduct a pose reproduction of the crystal structure ligand to ensure that we are able to successfully re-produce the low-energy pose that the ligand was crystalized in. If docking is not successful at this, then this system may not be suitable for further testing using DOCK since it is not able to produce a pose that is already known to be sampled by the ligand, based on observations from the crystal. For these reasons, pose reproduction is a worthwhile preliminary step in any docking experiments so that we can be more confident in our results if the system will be used for more complex docking protocols such as virtual screening or de novo design.
In the current implementation of DOCK6, a sampling success is defined as a pose RMSD that is < 2 Å relative to the crystal pose.
Rigid docking is the least computationally intensive algorithm in the DOCK program since it does not allow for dihedral rotations, or bond length perturbations. Since we are just focusing on pose reproduction (and since we know that the crystal structure ligand pose is of biological relevance)