Computer-Aided Drug Design Tutorials: Visualization of Electrostatic Potential Surfaces
Binding Pocket of Acetylcholinesterase
Acetylcholinesterase is an enzyme in the blood and central nervous system that breaks down neurotransmitter acetylcholine. Acetylcholine esterase inhibitors are used for the treatment of glaucoma, neuromuscular diseases, and memory loss in Alzheimer's patients. Follow these instructions to locate and analyze the substrate binding pocket in acetylcholinesterase that has been crystallized in complex with non-hydrolyzable substrate analog acetylthiocholine.
- Visit the PDB and search for "acetylthiocholine". Notice that over 30 structure hits were found but not all of these contain acetylthiocholine. Files 2C4H and 2C58 have structures with reasonable resolutions (2.15 and 2.30 Å, respectively). These structures were obtained from crystals grown in the presence of high (2C4H) and low (2C58) concentrations of acetylthiocholine, respectively.
- Select 2C4H first and examine this entry by clicking on the title line. Notice that this is a single-chain protein (Chain A), and that the bound acetylthiocholine carries an identifier AT3. Click on the View under 3D-ligand Interactions on the AT3 row and examine the active site in the Java visualizer. Notice that this structure contains two bound acetylthiocholine molecules in distinct locations. Could you tell from this structure which of the sites binds acetylcholine tighter?
- Start or reinitialize PyMOL
- Load the PDB structure 2C58 that contains acetylcholine only in the high-affinity site via PDB Loader Service Plugin.
- Display one of the bound non-hydrolyzable substrate analogs (residue name AT3) in the "ball-and-stick" representation keeping the coloring by element.
- Create an object that holds the binding pocket residues that are in near-contact (within 4.2 Ang) with the bound ligand AT3 by typing create pocket, (byres (resn at3 around 4.2))
- Show the binding pocket atoms and bonds in stick representation. To better mask the background, color the protein in a neutral color (e.g. color wheat, chain A and then re-display the binding pocket in colors by element (select C -> by element and pick your color combination from the right-hand task bar). Would you expect that such binding pocket has high affinity for positively charged small molecules such as acetylcholine?
Electrostatic Potential Surfaces
It seems surprising that the binding site for a small positively charged acetylcholine is so hydrophobic, and that there is not a single negatively charged group near acetylcholine. To understand ligand binding to acetylcholine better, we will now create an electrostatic potential map for this enzyme. In the process we see that advanced calculations do not always work well with unedited PDB files, and we will learn a bit about editing PDB files.
Electrostatic potential maps for proteins are typically calculated by numerically solving Poisson-Boltzmann equation based on atomic point charges derived from molecular mechanics force fields (in the case of ultra-high-resolution X-ray structures, the charges can be calculated from the observed electron density as well). The charge assignment requires complete residues with hydrogen atoms and involves an assessment of side-chain pKa's. PyMOL can automatically perform some of these tasks; alternatively PSB2PQR server can be used to generate suitable input files for such calculations. PyMOL uses the APBS Plugin, that invokes a freely available Adaptive Poisson-Boltzmann Solver for the generation of electrostatic potential maps.
- Save the PDB file object 2C58 in your directory using Save Molecule under the File menu.
- Start APBS Tools from Plugins menu. Hit on APBS Location tab. Notice that the path to executable is /usr/bin/abps on workstations. If you have installed APBS in another directory, you must change this line appropriately (e.g. there is a slightly newer version of APBS at /home/chem162/Software/apbs_new/bin/apbs.
- Hit Set Grid, then Run APBS. At this point PyMOL complains a lot. Some of the informative complaints are "selection cannot span more than one object." and "# complete the carboxy terminus". Looks like there are some issues with the PDB file that we must fix before using APBS. Close PyMOL for now.
- Examine the PDB file (bluefish 2C58.pdb). There are a few issues with the PDB file that was deposited with PDB. Scroll past the headers and look at the lines that describe residue 19. Notice that the PDB file describes two alternative conformations for this residue. This is good if you are interested about protein flexibility, but electrostatic potentials cannot be calculated with files that describe more than one structure.
- What needs to be done is to change "AARG" into "ARG" and eliminate "BARG" lines. You could do this manually but it gets tedious after some time, especially if working with high-resolution structures that contain many such resolved flexible residues. If something gets tedious, one should write a computer program to perform this tedious job. Very-high-level languages, such as Python offer relatively easy ways to write computer programs that perform simple repetitive tasks. Download a small program pdb_prep.py into your directory and examine the program by typing more pdb_prep.py. Can you deduce what this program does?
- Run the pdb_prep.py program on your PDB file. By default, downloaded programs under Linux are not executable and you have to set the permissions to 755 with a command chmod 755 pdb_prep.py. To run the program, type ./pdb_prep.py 2C58.pdb (In Windows, use command line and type pdb_prep.py 2C58.pdb after installing Python; MacOS X comes with Python installed). Examine the two output files that it generated. The program nicely separated the protein and the ligand parts. This was necessary because the APBS plugin knows only how to calculate the electrostatic potential based on the protein residues for which partial charges can be looked up from a table.
- There is one last problem with the file. Recall that PyMOL instructed us to complete the carboxy terminus. Open the file with protein's structure with text editor vi 2C58_prot.pdb, hit Enter, type G to move to the end of the file. The last residue clearly lacks C-terminus but instead of trying to build it, we can just truncate the protein. Hit DD to delete the last line, then ZZ to save the modified file. While cryptic, vi gets the job done fast. For people less inclined to edit their PDB files, or otherwise strongly opposed to reinventing the wheel, the online PDB2PQR servers may be able to eliminate most common problems when working with APBS.
- Read the modified protein file into PyMOL by typing pymol 2C58_prot.pdb.
- Open the APBS plugin; you may change to the newer version of APBS if you like. Hit Set Grid, then Run APBS. The calculation will take about a minute, and you can follow the progress of calculation on the terminal that was used to launch PyMOL.
- Hit the Visualization tab in APBS window and Update view. Change the low and high potential values to -3 and +3, respectively, and hit Show. Wait for the surface to be generated.
- Rotate the molecule slowly and examine the electrostatic potential. You can drag with the mouse middle button on the scale bar while holding down the Cntr key to change the potential range. Identify a cavity that is negatively charged (red) and thus could attract acetylcholinesterase.
- Open the file 2C58_lign.pdb and display the ligand AT3 as sticks. Confirm that the ligand is indeed bound to the pocket that has strong negative potential. You can make the surface semi-transparent by using a command set transparency, 0.3 so that the binding pocket residues are partially visible through the surface.
- Save a picture of the view on the screen using Save Image under File menu.
- Save the completed work using Save Session As under File so that you can return to this point if you need to.