Preparing a PDB File

Figure 1: Schematic view of the ligand-binding domain from the vitamin D receptor (PDB file 1IE9). The crystallographic waters are shown as small spheres and the bound ligand is shown as a CPK model. HO H 3 C CH 3 H C H 3 CH 2 OH OH CH 3 Figure 2: 1α,25-dihydroxyvitamin D 3, the metabolized form of vitamin D. It is named VDX in this exercise. Preparing a PDB File Load a PDB file directly from the Protein Data Bank, Split the molecule, Edit non-protein components, Combine components, Rename objects using the Hierarchy View, and Prepare and run an energy calculation. The Protein Data Bank (PDB) is possibly the world s leading public source of three-dimensional data for biological molecules (1). As of July 2006, over 37,000 entries could be found in the PDB. Hundreds more are being added every month. Both X-ray diffraction and other solid-state techniques account for the majority of the structures. However, over 5500 NMR structures are also available. These deposited structures include proteins, peptides, nucleic acids, carbohydrates, and complexes of these molecules. As a first step in a modeling project, many researchers look in the PDB to find available structures related to their project. Preparation of these molecules for work in the Discovery Studio environment is a critical process to your modeling efforts. In the following steps, we will load a PDB file for the ligand-binding domain of the vitamin D receptor (VDR) in complex with a ligand (named VDX in this exercise). The file is 1IE9 as reported by Tocchini-Valentini et al. (5). The vitamin D receptor is responsible for the expression of a variety of genes including calcium metabolism, bone formation, and cell growth and differentiation (2). Understanding VDR conformational changes resulting from interactions with bound ligands may help to identify and treat persons at risk for disorders such as osteoporosis, breast cancer, or prostate cancer. In this exercise, you will retrieve the PDB file and prepare it for an energy calculation such as an energy minimization or molecular dynamics simulation. You will learn in the lesson how to: Seq1-2

1. Load the PDB file We must have Discovery Studio running to start the exercise. Launch the Discovery Studio client if it not already running. If Discovery Studio is already running, from the Windows menu, select the Close All command. If you are prompted to save any molecules or data, select No. Now, we will load the PDB file directly from the RCSB through the Discovery Studio interface. Note: The File Open URL command only will obtain files through a network connection to a PDB server. An error is returned if the connection is not possible. Check with your instructor or system administrator if a connection is available from your workshop location. From the File pulldown menu, select the Open URL command. In the dialog box, the URL should refer to the RCSB site. Replace the last four characters of the URL with 1ie9. Click the Open button. If a connection to the PDB cannot be made, the required file is available in the DataFiles directory as 1IE9.pdb. The structure displayed in the 3D Window is the crystal structure of the ligand-binding domain of the vitamin D receptor with a bound ligand and 225 crystallographic waters. From the View menu, select the Hierarchy command. Note the distribution of components in the Hierarchy View. We will rearrange the components to facilitate working with the structure. 2. Remove the unit cell The crystallographic unit cell often has little meaning in molecular mechanics calculations. Sometimes it is just easier to remove the unit cell. From the Structure menu, select the Crystal Cell pullright menu. Choose the Remove Cell command. Note the change in the Hierarchy View as the unit cell information is removed. Expand the object 1ie9 to verify that the ligand and water molecules are still present. 3. Split the components of the molecular system This PDB file is composed of many parts. We can view the components through the Hierarchy Window explored in other exercises. However, we can actually separate the components of the PDB file quicker with the Split command. Access the Tools Explorer. Then, access the Protein Reports and Utilities tool. Find the Split tool group. Click the All command. Three options are listed and will be described in a moment. Note the change in the Hierarchy View. The three components of the PDB file have been segregated into separate objects and renamed. The three components are the protein (1ie9_A), the ligand (1ie9_NonProtein), and the Seq1-3

crystallographic water (1ie9_Water). The three options for the Split command allow some flexibility in how to manipulate objects. All Protein Non-Protein Splits out all chains and non-protein substructures into separate objects in the Hierarchy View and lists each amino acid sequence as a separate sequence in the Sequence View. Splits out all protein chain substructures into separate objects in the Hierarchy View and lists each amino acid sequence as a separate sequence in the Sequence View. Splits out ligands and other non-protein substructures such as waters into separate objects in the Hierarchy View and lists any non-protein polypeptide sequences as separate sequences in the Sequence View. 4. Remove the water For this exercise, we will remove the crystallographic water molecules. There are several ways that the waters could be removed. In this case, as we split the system and produced a new hierarchical arrangement, we will use the Hierarchy View. In the Hierarchy View, select the entry 1ie9_Water. The entry should be selected in both the Hierarchy View and the 3D Window. Click the Delete key on the keyboard. The 225 water molecules are removed from both the Hierarchy View and the 3D Window. 5. Rename components To simplify the object names, we may rename them. First we will rename the overall system. In the Hierarchy View, select the entry 1ie9_A. Right click with the mouse and select the command Attributes of 1ie9_A. In the Molecule Attributes dialog, select the Name cell. Change the entry to Complex Click the OK button. Now we will rename the protein chain Expand the entry for Complex. Select the entry A. Right click with the mouse and select the command Attributes of A. In the Molecule Attributes dialog, select the Name cell. Change the entry to Receptor. Click OK. Now, we will group the ligand VDX with the receptor. While holding down the left mouse button in the Hierarchy View, drag the entry for 1ie9_NonProtein into the entry Complex. The resulting hierarchical arrangement should appear as in Figure 3. Note the ligand has been renamed <Chain>. This is the name of the molecule that we would have seen under the object 1ie9_NonProtein. Seq1-4

Figure 3: Hierarchy view with the merged objects Finally we will rename the ligand. Now select in the Hierarchy View the entry <Chain>. Right click with the mouse and select the command Attributes of Chain. Enter for the object name VDX. Click the OK button. The final Hierarchy View should appear as in Figure 4. Figure 4: Hierarchy window showing the merged and rename objects 6. Clean the protein Discovery Studio can automate many of the tasks required to properly prepare a protein for an energy calculation. Before performing the clean operation, we can specify what operations to conduct through the Preferences dialog. From the Edit menu, select the Preferences command. In the Preferences dialog, expand the Protein Utilities page. Select the Clean Protein page. At this point, we have several options that can be set. These options will address common problems that may be present in PDB files. For example, X-ray crystal structures will typically not have hydrogen atoms, so these must be added. Also, the chain ends must be set with the correct chemistry and any missing atoms in residues must be placed. The options are described below. Correct Problems Desired ph NonStandard Names Disorder (Retain One Set) Incomplete Residues Checks whether atom names conform to the standard names and corrects them if necessary. Checks for disordered atoms and retains only the first set. Adds missing side chain atoms to amino acids. This operation will not fill in missing loop regions. Allows the protonation state of the ionizable amino acids and the termini to be controlled using the standard pk a values. The protonation state is adjusted after any modifications of the hydrogens and termini by the following options. Seq1-5

Figure 5: Clean Protein preferences panel Hydrogens Termini If the Modify Hydrogens option is checked, hydrogens will be added as needed. All Hydrogens: All hydrogen atoms are added. Polar Hydrogens: Only those hydrogen atoms that could be involved in hydrogen bonds will be added. No Hydrogens: No hydrogen atoms are added. If the Modify Termini option is checked, then termini will be added or removed as indicated. Fix Connectivity and Bond Orders Ensures that amino acids have the correct bond order. This option will have no effect on nonstandard amino acids or ligands. We will set the Clean Protein preferences now. Turn on preferences as shown in Figure 5 on the previous page. Turn on all toggles in the Correct Options section. Toggle on Modify Hydrogens and All. Toggle on Modify Termini and Add. Click the OK button in the Preferences dialog and return to the 3D window. Now, we may actually clean the protein. In the Hierarchy View, select the entry Receptor. Access the Tools Explorer. Then access the Protein Reports and Utilities tools. Find the Split tool group. Click the Clean command. Seq1-6

After a few moments, the protein structure that is present in the 3D Window changes. Most notably, hydrogen atoms are now added to the structure. Other operations have also been performed such as completing a missing histidine residue at position 377 and correcting atom names. If you look at the residue listing in the Hierarchy View, you will notice a jump in the residue numbering from 164 to 216. The protein crystallized is a fusion protein with a variable insertion domain removed (4). The numbering of the residues in the PDB thus jumps but does have a proper peptide bond filling the gap. 7. Edit the inhibitor structure Now let us closely examine the inhibitor molecule. Select VDX in the Hierarchy View. With VDX selected, go to the View menu and select the Visibility pullright menu. Choose the Show Only command. The protein disappears from the 3D Window leaving only the ligand. Click the Fit to Screen button. Deselect everything. Notice that all bonds are set as single bonds. PDB files often lack information regarding the bond order for ligands. In order to properly model any protein-ligand interactions, it is necessary to have the correct bond orders and the correct number of hydrogen atoms. The Clean operation performed earlier would not modify or clean the ligand. The Clean function only operates on standard residues such as the standard amino acids. For ligands and other nonstandard residues, we must manually correct the structures. First, let us label the molecule. From the Structure menu, select the Labels pullright. Then select the Add command. For the Object, select Atom. For Attribute, select the Name option. Click the OK button. Now we must form three double bonds to have the bond order agree with that displayed in Figure 2. Enlarge the ligand as necessary to see the six-member ring clearly. Note the three atoms coming off the ring. Two of those atoms are oxygen atoms (red in color). The third atom is a carbon (labeled C19) and must form a double bond as shown in Figure 2 on the first page. Using the Chemistry Bond command, form a double bond between C10 and C19. Now form two double bonds between the two rings. One double bond is to be formed between atoms C5 and C6 and another between C7-C8. Once the bond order has been corrected, it is necessary to add hydrogen atoms. From under the Chemistry pulldown menu, select the Hydrogens pullright menu. Choose the Add command. From the Structures menu, select the Labels pullright. Then select the Remove command. Seq1-7

8. Type the molecules With the proper chemistry set in both the protein and ligand we may perform the atom typing. First, let us ensure that we have the desired settings. From the Edit menu, select the Preferences command. Access the Forcefield panel. Click the Reset button. This will reset the typing preferences to the default setting. Most notably, the Momany and Rone CHARMm force field will be used (3) with automatic typing and Momany-Rone charging rules. Click the OK button. Now redisplay the protein with the ligand. Under the View menu, select the Visibility pullright menu. Choose the Show All command. Click the Fit to Screen button. Now we may perform the atom typing. Ensure nothing is selected in the 3D Window. Access the Tools Explorer. Then access the Forcefield tools. Ensure the Forcefield: parameter is set to CHARMm. Click the Apply Forcefield command. The Status changes to read the 1IE9 is now properly typed. The name in the status line refers to the 3D Window, thus both the receptor and ligand have been typed. Both molecules are now properly typed and are ready for an energy calculation, 9. Preparing for an energy calculation For this exercise, we will simply calculate a point energy for the system. That is, the current conformational energy of the system will be determined. If this operation can be performed without error, we probably will be able to perform other more extensive calculations. Access the Protocol Explorer. Select the Simulation group. Double click on the Energy protocol. In the Parameter Explorer, click on Typed Input Molecule. Ensure the entry for the parameter value is 1ie9:Complex. Leave the other settings to their default values. Click the Run button. The run will only take a few seconds to run. When the Job Completed dialog appears, click OK. Note that in the Job Explorer in the Details column the total energy of the system has been reported. As we have an energy value, we are confident the peptide chain and inhibitor were successfully cleaned and typed for CHARMm. 10. Saving the new complex We will save this single protein-single ligand complex for later use. Seq1-8

Ensure the 3D Window is active. Copyright 2006, Accelrys Software Inc. All rights reserved. From the File pulldown menu, select the Save As command. Click the Desktop icon to save the file to your Desktop. For File Name, enter the name 1ie9_vdx. Ensure the File of type: parameter is set to Viewer (*.msv). Click the Save button. 11. Complete the lesson Close all the windows in Discovery Studio. From the Window menu, select the Close All command. Alternatively, you could simply exit Discovery Studio. If asked to save any data, select No. With this lesson, we have taken a PDB file composed of various molecular entities. We have broken apart these molecular components and regrouped them as needed. Further, we have seen that, while we can clean automatically polypeptide chains, we must edit PDB heteroatoms to attain the correct chemistry. Once we have the desired molecular system with the correct chemistry we may perform the desired calculations. The proper preparation and organization of molecular systems is critical to obtain reasonable results. In later lessons and workshops, we will examine how to better evaluate the protein structure and especially protein-ligand complexes. Seq1-9