MOE/DivCon Tutorial for 1AZM Case Study
As discussed in a recent case study, correct protonation determination is a key goal in structure based drug discovery. Unfortunately, very often we are unable to resolve protons at the refinement resolutions we deal with on a day-to-day basis. Therefore, we need a tool to correctly determine protonation state and explore how well it agrees with experiment. With QM-based refinement available in Phenix/DivCon, we are able to marry the best computational methods with the best experimental methods and inform the protonation question. In the steps below, we will use the SVL (Scientific Vector Language) plugin provided by QuantumBio in order to use MOE to prepare the structure for refinement.
The benefits of using MOE are two fold: (1) with the graphical user interface we can double check the protonation states of our species and (2) the Protonate3D tool available within the interface does a good job of protonating most biomolecular structures. However, as with any perception based method – such as Protonate3D – it will only perform as well as the input structure allows. For this reason, it is always best to double check your work!
The environment must be properly prepared for the simulation. This includes sourcing the Phenix environment file prior to to running MOE. Since we will be using qbmoe to run MOE, we do not need to also source the QuantumBio environment file. We will also need to create an empty directory and place our starting PDB and structure factor file within the directory. These files are available at PDB.org and may be downloaded and placed within this 1AZM directory.
% mkdir 1AZM ; cd 1AZM
% cp /path/to/repository/1AZM.pdb .
% cp /path/to/repository/1azm-sf.cif .
% source /path/to/phenix-1.9-xxxx/phenix_env.sh
Within the empty MOE graphical user interface window, choose File->Open and open the 1AZM.pdb file. You should also choose to open it with PBC turned on in order to make sure that protonation and any other cleanup is performed with symmetry in mind. You may also choose to include Alternates; however, any alternates within the QM region will be stripped out during the refinement.
When the structure appears on the screen, it is ready to be cleaned up, protonated, and refined. For the sake of discussion you may want to center the Zn++ on the screen and do some visual tuning in order to follow along and observe the protonation states we will manipulate in the subsequent steps. For example, see the following screenshot:
The Refinement Preparation plugin is accessed in the QuantumBio menu at the top of the screen. If you do not see this menu then please double check your installation of the DivCon Discovery Suite and come back to continue.
Choose the QuantumBio->Refinement Preparation… menu item and the following widget will appear on the screen.
Proceed to Browse… and enter the 1azm-sf.cif MTZ/SF file copied from pdb.org above. Once this is ready, we will run the MOE Structure Preparation tool to clean up any problems and run Protonate3D.
When the MOE Structure Preparation tool appears, you may choose to perform any cleanup you wish. For more information, please see the MOE documentation. For this tutorial we went ahead and simply clicked the Correct button on the panel then clicked the Protonate3D button. We generally set the pH to 7.4 as we have found this setting to be less likely to over-protonate HIS residues and the like. Once you have chosen the various settings on the screen, click OK. On the Structure Preparation dialog box, press Close.
Protonate3D will run for some time until it has performed all of the various calculations. Once the protonation is complete, find the HIS94, the GLN92, the Zn++, and the AZM ligand. Be sure to also show all of the protons. You should see that HIS94 has been given a negative charge and is hydrogen-bound to the GLN92 (the latter has been “flipped” vs the original PDB structure). Since our goal is to explore two protonation states of this HIS94, we will save this structure for refinement.
To move forward, once you have pressed the Close button in the Structure Preparation dialog box, the Phenix/DivCon Preparation dialog box should reappear. Before we generate the batch file, it is always best to check to make sure that we are including the correct atoms within the QM Region and the Buffer region. Click the question mark (?) buttons to double check. It should become evident that with the default setting of only the ligand AZM chosen as the Core Residue that some of the Zn++ coordination sphere is missing from the QM region. Therefore, clear the selection from the last ? button picked by clicking the Select->Clear menu item. Now select the Zn++ atom on the screen and then click the Retrieve Selection button within the Phenix/DivCon Preparation widget. The ZN residue will then appear within the Core Residues selection box, and clicking the ? buttons after each region size will show that the coordination sphere is included.
Take note of the expected memory size, manipulate the number of cores as desired, etc. Finally, click the Batch… button along the bottom of the widget to generate the batch file (we call it 1AZM_batch_NegHIS94.sh).
Now use standard MOE tools to change the protonation state of the HIS94 to protonate the nitrogen. This requires using the MOE Builder to change the bond orders of both of the N=C and C-N bonds to N-C and C=N to make sure that the protonation states are correct.
Next, the “flip state” of the GLN92 must be manipulated in order to return it to its original state. This can be accomplished by selecting the entire GLN92 residue, and using Protonate3D on just those selected atoms. The Protonate3D dialog box can be configured in this way and the GLN92 will flip back:
Once the structure is protonated and prepared as required, again select the QuantumBio->Refinement Preparation… menu item. This time, since the structure is ready to go, we only need to choose the same structure factor file as last time, select and Retrieve Selection for the Zn++, and then click the Batch… button. Accept the warnings as presented, and save the file as 1AZM_batch_ProtonatedHIS.sh.
These two files (1AZM_batch_NegHIS94.sh and 1AZM_batch_ProtonatedHIS.sh) are fully self contained files which package the input pdb and mtz files along with any other files required (other than the Phenix/DivCon software). You can move them to your cluster and they have all of the input available. They will source the same Phenix environment you had used at the top of this page, and they will prepare an identical environment. Therefore, if your cluster has a different directory environment than your workstation, you should edit the batch scripts within a text editor and change the files where the Phenix and DivCon environments are defined at the top of the script.
Once you have prepared the files, you can simply run them using the following command line options:
mkdir NegHIS_run ; cd NegHIS_run cp ../1AZM_batch_NegHIS94.sh . ./1AZM_batch_NegHIS94.sh >& SCREEN & cd .. mkdir ProtonatedHIS_run ; cd ProtonatedHIS_run cp ../1AZM_batch_ProtonatedHIS.sh . ./1AZM_batch_ProtonatedHIS.sh >& SCREEN &