BCHM 32200

Week 14

Computational Modeling:
Ligand Docking, in silico Mutagenesis, and Inhibitor Optimization

BACKGROUND

See lecture slides for basic information on molecular docking.

Computational modeling tools can be an efficient and valuable approach to studying the interaction of
small molecules with proteins that have well-defined atomic resolution structures (e.g. from X-ray
crystallography). Applications include the identification and design of inhibitors in the drug discovery
field, as well as mechanistic understanding of how enzymes recognize and act on substrates.

In todays exploratory lab, you will use the MOE software you worked with previously to create
homology-based structural models for your enzymes. Using those homology models you will perform
the following exercises:
1. Docking of a substrate peptide from the hCdc114B crystal structure into the active site of your
enzyme and in silico mutagenesis to understand the enzymes specificity for phosphoserine.
2. Docking of the inhibitor studied last week into your enzymes active site
3. Optimization of the inhibitor to create modified structures with enhanced binding potential
and specificity.

How does this relate to your overall projects? First, understanding how optimal substrates bind
in the active site provides useful information for the design of inhibitors, many of which act by
mimicking substrates.
Second, the identification of inhibitors (like the 5224815 compound you used last week) through
methods like high throughput screening of chemical libraries is usually only the first step in development
of a useful drug (or pesticide). Additional rounds of modification and optimization of various properties
of a compound are typically necessary to create something that is maximally effective, cost-effective to
produce, stable, safe, etc. Your ligand docking and optimization exercises today should give you a brief
glimpse of what this process is like once a lead compound has been identified for a given target enzyme.
The new compounds you design would then need to be synthesized and tested for improved
performance (e.g. lower Ki values, activity towards live fungal cells, etc).
INSTRUCTIONS

A. Substrate peptide docking in the homology model active site.

1. Load the structure of the human Cdc14b phosphatase in complex with a peptide (1OHE.pdb)
using the PDB button on the left in the Open File window. Execute the Protonate 3D
functionality first (Main Menu Compute Prepare Protonate 3D use default
settings) this predicts positions for all hydrogen atoms and adds them to the model.
Delete the last two chains in the Sequence Editor (SE) window, which consist of just water
molecules.

2. Using the right menu options, portray the structure as secondary structure elements only
(ribbons and arrows). You will have to Select All Atoms and then Hide side chains, backbone,
solvent, etc.

3. Next, open the structural model .pdb file of your Cdc14 enzyme from your OneDrive folder and
run the Protonate 3D function as well. Get your model portrayed as secondary structure as
well and color it differently than the human structure. Remember, you can use the SE
window to select individual chains when you have multiple molecules present.

4. Perform an alignment of the two Cdc14 sequences and then Superpose the two structures
(Main Menu Protein Align/Superpose and make sure in the Align window you select
both sequences in the top window and that Sequence and Structural Alignment is selected
in the Action dropdown menu at the bottom. Click on the Align button first. Then click on
the Superpose button. The two structures should be overlaid very nicely now.

5. Next, select the phosphopeptide substrate by double clicking on the B chain of the hCdc14B
structure in the SE window and click on Show Selected from the right menu bar. Display
the peptide in ball and stick form and color by element if it doesnt already appear that way.
With the peptide atoms still selected, designate it as the ligand (Right Menu Ligand
Choose Ligand). Zoom into the binding site by clicking Site View on the right menu. You
should now see the short peptide molecule and its binding site in the middle of the screen.
If H-bond contacts are shown deselect that option for now on the bottom task bar Contacts
menu.

6. Go back to SE and select and delete the hCdc14B protein chain. Now you are left with just the
short peptide and your own Cdc14 structural model. The peptide should be lined up with
the binding site on your Cdc14 protein since the two structures were superposed. The
position of the peptide will define the ligand binding site on your protein.
7. Convert your protein structure back to thin stick representation using the Select All function
from the right menu along with the Atom and Ribbon menus on the bottom task bar. Then,
set the substrate binding site to surface representation (Right Menu Surface Receptor)
and then make it semi-transparent using the slider bars (Right Menu Surface Surfaces
and Maps)

8. Save your .moe file to your desktop with your group number and the name Week 12 in the title
(for example group1_week12_active site.moe).

9. We will now perform two different docking experiments to see if the peptide is predicted to
bind the same to your protein as it does to the hCdc14B protein. Open the docking menu
(Main Menu Compute Dock). Set the following parameters:
Receptor: MOE in left dropdown menu; Receptor and Solvent in right menu
Ligand: Ligand atoms
Placement:
o Method: Triangle Matcher
o Score: London dG
Refinement:
o Method: Rigid Receptor
o Score: GBVI/WSA dG
Poses: leave as default
Any other settings, leave as default
Direct the output file to your desktop so it can be found easily later and moved to
your OneDrive folders. Name the file Cdc14_pepdock_rigid_Group#.mdb

10. Execute the docking protocol by hitting Run. The Database Viewer Window (DBW) will open
up and you will see in real time the different ligand poses being tested in the models
binding site. Once the run is finished, the top 20 poses will be reported in the DBW by
dragging the bottom right corner of a cell in the first column to make it bigger you can see
the individual pose structures. The score represents the estimated free energy of binding
remember the more negative the value, the stronger the interaction. The hits will be ranked
by this score.

11. You can then view each of the 20 docking results in the Database Browser (From DBW go to File
Browse). You can minimize the DBW and move the browse window and your structure so
you can see the binding site clearly. Click through all docking poses and compare them to
the ligand pose from the crystal structure. Take note of the poses that most closely
 resemble the actual pose from the crystal structure. Save .jpg images of the two poses that
most closely match the crystal structure (MAKE SURE YOUR GROUP # IS IN THE FILE NAME).
Make sure you understand the structure of the ligand and can distinguish the different
functional groups on the short peptide (e.g. the phosphate group, the proline side chain,
etc). Also save a screenshot of your DBW showing the top 5 poses with their scores and
other values.

12. Now close the browse and DBW windows and repeat the entire docking procedure again with
the exact same settings EXCEPT select Induced Fit this time from the Protocols dropdown
menu at the top of the Dock window. Also, provide the new database file name
Cdc14_pepdock_indfit_Group#.mdb. Repeat the analyses and save the best two poses as
.jpg files and the DBW screenshot of the top 5 poses as well.

13. Finally, save an image of the best docked peptide pose WITHOUT the original peptide structure
present (Select and Hide the ligand before running through the Browse function), and create
and save a 2D interaction map of your docked peptide, making sure your group # is in the
filenames.

B. in silico mutagenesis
You discovered a couple weeks ago that Cdc14 enzymes all seem to share the unusual ability to
strictly distinguish between phosphoserine and phosphothreonine. How are they able to do this?
One suggestion is provided in the 2012 paper you were provided with previously (Bremmer et al. J.
Biol Chem) from evaluation of the hCdc14B structure bound to the short phosphoserine peptide you
just performed docking with. Look at that paper and see if you can find what the proposed
explanation for specificity is. Then, use MOE and the hCdc14B structure (1OHE) to test that
hypothesis by modeling by following these instructions.

1. Close all open files In MOE, then load the 1OHE structure again and get to the same binding site
view with the surface representation of the protein and ball and stick of the peptide substrate
(follow instructions from section above). Find the phosphoserine of the substrate and then
remove the surface representation then so you can see contacts between substrate and specific
amino acids in the protein.

2. Select the peptide substrate and click Show Hydrogens from the right side menu bar. From
the bottom task bar Contacts menu check the clashes box. Steric clashes will be indicated by
orange disks of varying sizes reflecting the extent of the clash. Save a .jpg file of this native
structure.
 3. Now, select just the phosphorylated serine residue in the substrate in SE. Then click on Protein
Protein Builder. Click the more button in the Protein Builder window that pops up and then
select the botton for Tpo, which will replace the phosphoserine with a phosphothreonine.

4. Do you see any steric clashes appear that did not exist before? If so, determine what amino acid
residues are involved in the steric clash with the pThr side chain. Label those amino acids and
save the new model as a .jpg file.

5. Predict and make a mutation in the enzyme active site that could relieve the steric clash with
phosphothreonine peptide. You can easily make mutations in your structure by going to SE,
selecting the residue to change, opening Protein Protein Builder and selecting the amino acid
you want to replace it with. Are you able to reduce the steric clash? Are the results of this
exercise consistent with what is proposed in the paper? Save this model of the mutant enzyme
also as a .jpg file. Save your .moe session for the mutagenesis exercise as well.

C. Inhibitor 5224815 docking in the homology model and human Cdc14B active sites.

1. Close the open MOE session and start over once again by loading the human Cdc14B 1OHE
structure and run through the same processing steps described in section A to get it ready.
You can display the protein in sticks representation.

2. Again, set the short peptide as the ligand, click on Site View, and make a semi-transparent
surface representation of the binding site.

3. Now, click on the Builder button on the right side menu bar. Go to the text file for Cdc14
inhibitor structures in the main OneDrive Data folder. Copy the structure code for the first
inhibitor, which is the one you characterized last week. Paste the code into the window that
says SMILES and click on the happy face button. Click yes when it asks you if you want to
create a new molecule. The structure of the inhibitor should now appear in or near the
binding site with the bound peptide.

4. Give the inhibitor molecule a name. From the right side menu bar select System click on grey
bar and then the open white corner at the bottom right of the large name cell and enter the
name 5224815 under tag.

5. From the SE window, delete the peptide ligand from the hCdc14B structure, select your inhibitor
molecule, and designate it as the ligand. Now, conduct the induced fit docking run again as
you did above with the same settings.
 6. Conduct the same analyses and save the same images as you did for the peptide docking above.
An important question is: do you see a common orientation for the inhibitor in the binding
site across the top scoring poses? Are other groups getting a similar orientation?

7. Using the Browse function again, display the best inhibitor pose only. You can hide the original
inhibitor molecule. Display hydrogen bonds using the bottom task bar Contacts menu and
figure out what hydrogen bond contacts are predicted for binding of this inhibitor to your
enzyme (what amino acid residues?) You can label amino acids using the bottom task bar
Atoms menu.

8. Color the receptor surface according to lipophilicity. See if you can predict hydrophobic
interactions between binding site amino acids and parts of the inhibitor molecule.

9. Finally, create a 2D interaction map for the docked inhibitor by selecting Ligand Ligand
Interactions from the right side menu bar. Make sure you select the docked inhibitor and
the adjusted receptor structure in the 2D interactions window and apply these changes.
Note the hydrogen bond contacts predicted for the interaction. Save the 2D interaction map
image.

10. Now, load your homology model structure and perform the alignment and superposition with
the hCdc14B structure, delete the hCdc14B structure and run the docking exercises with the
5224815 inhibitor again, saving the exact same images and information as for the hCdc14B
enzyme. Important save the .moe session of your enzyme docked with the top inhibitor
pose for the next section.

D. Inhibitor Optimization

1. Open the MOE session from part C containing your homology model with the docked 5224815
inhibitor molecule.

2. Make sure the inhibitor is set as the ligand (RHS| Ligand Choose Ligand).

3. Execute the Site View command (RHS | Site View).

4. Display the receptor atoms (RHS | Select Receptor) in a color of your choice (Footer | Atoms
pick a color). Turn off the ribbon display (Footer | Ribbon).

5. Change the force field from MMFF94x to Amber12:EHT (Footer | Potential Setup in left corner).

6. Make sure you have a 2D ligand interaction diagram saved for this inhibitor pose. If not, do it
now (RHS | Ligand Ligand Interactions).
 7. Turn on the hydrogen bond and steric clash displays (Footer | Contacts). Save an image of the
inhibitor bound to the enzyme active site with the bonds and clashes displayed.

8. Display the calculated ligand properties (RHS | Ligand Properties). First, minimize the
conformational energy of the ligand/enzyme complex (RHS | Minimize). The calculated
property values (upper left corner of the main window) are:

Free Energy of binding (solvation): __________ Kcal/mol

Docking Score (affinity): __________ Kcal/mol

9. Based on the interaction diagram and the electrostatic surface map (and perhaps R-vectors, if
you plan on adding groups), make a hypothesis about TWO structural modifications that you
believe will enhance the affinity of the ligand for the receptor. Note that you do not
necessarily have to ADD groups modifications of existing groups are equally good, if not
even better.

hypothesis 1: ________________________________

hypothesis 2: ________________________________

10. Create the new molecules with the Builder (RHS) functionality and minimize the new structures
conformational energy (again with the Amber Force Field) to obtain predicted binding
affinities.

hypothesis 1:

Free Energy of binding (solvation): __________ Kcal/mol

Docking Score (affinity): __________ Kcal/mol

hypothesis 2:

Free Energy of binding (solvation): __________ Kcal/mol

Docking Score (affinity): __________ Kcal/mol

Try additional modifications if you think you can improve the predicted binding affinity even more. See if
you can take advantage of known substrate binding sites like the +1 Pro pocket. Save images of any of
the new molecules bound to the enzyme active site with H-bonds and clashes shown AND save 2D
interaction diagrams. How much have you improved the predicted binding affinity with your
modifications?
Make sure the following files are saved in your OneDrive Week 14 folder:

Part A: .moe file after alignment of hCdc14B and homology model with active site shown with Site View

2 peptide docking result .mdb files (rigid and induced fit protocols)

4 .jpg files (2 per docking exercise) of most similar binding poses compared to crystal structure.

2 screenshots of DBW window with top 5 peptide binding poses shown (1 for each docking run)

1 .jpg file of final docked peptide pose without the original peptide shown

2D interaction diagram of final docked peptide pose

Part B: .jpg files comparing phosphoserine and phosphothreonine peptides binding to enzyme as well
as the mutant enzyme bound to the phosphothreonine peptide with clashes shown and
residues labeled.

.moe file of the mutagenesis session.

Part C: .mdb, .jpg, and 2D interaction diagram files for inhibitor molecule docked to both hCdc14B and
your enzyme (top 2 poses each)

.moe file of your enzyme docked with the top inhibitor pose

Part D: 2D interaction diagrams and .jpg images of original inhibitor and at least 2 different modified
inhibitor molecule bound to active site with H-bonds and steric clashes labeled.