[ Assign Helices and Strands | Protein Health | Local Flexibility | Protein Interface by ODA | icmPocketFinder ]

3.6.1 Assign Helices and Strands

Theory

The Assign helices and Strands option will manually reassign secondary structure to a protein structure. This command does not change the geometry of the model, it only formally assigns secondary structure symbols to residues. f the secondary structure string is not specified, apply ICM modification of the DSSP algorithm of automatic secondary structure assignment (Kabsch and Sander, 1983) based on the observed pattern of hydrogen bonds in a three dimensional structure. The DSSP algorithm in its original form overassigns the helical regions. For example, in the structure of T4 lysozyme (PDB code 103l ) DSSP assigns to one helix the whole region a_/93:112 which actually consists of two helices a_/93:105 and a_/108:112 forming a sharp angle of 64 degrees. ICM employs a modified algorithm which patches the above problem of the original DSSP algorithm. Assigned secondary structure types are the following: "H" - alpha helix, "G" - 3/10 helix, "I" - pi helix, "E" - beta strand, "B" - beta-bridge, "_" or "C" - coil.

To assign secondary structure:

• Load the pdb structure (File/Open or PDB Search}
• Select the structure. You can do this by double clicking on the name of the structure in the ICM Workspace (a selection is highlighted blue in the ICM Workspace and green crosses in the graphical display) or you can use the right-click button and drag it over the whole structure in the graphical display.
• Tools/3D Predict/Assign helices and Strands

3.6.2 Protein Health

Theory

The protein health option calculates the energy strain of a structure in ICM. It is generally a good idea to investigate the energy strain of any protein structure before undertaking such processes as docking. It is also essential to use this tool after making a model (see Molecular Modeling) to identify strained regions within your model and then some optimization procedure can be undertaken to rectify the problems.

The protein health option calculates the relative energy of each residue for a selection and colors the selected residues by strain.

This macro uses statistics obtained in the following paper Maiorov, V.N. and Abagyan, R.A. (1998) Energy strain in three-dimensional protein structures Folding and Design, 3 , 259-269.

To use the Protein Health option your structure must be converted into an ICM object (see Converting to ICM Object)

Next, make a selection of which residues you wish to analyze (see Making Selections).

• Tools/3D Predict/Protein Health and a window as shown below will be displayed.

• The scale of the coloring can be changed by altering the value within the trimEnergy data entry box.
• Click OK and the structure will be colored according to energy strain (red - high) and a table of residue energy will be displayed in a table.
• To reactivate the screen click the Go button in the bottom left hand corner of the GUI display.

The Protein Health option returns a table of energies for each amino acid in the selection:

The Protein Health option returns a plot of energies for each amino acid in the selection:

3.6.3 Local Flexibility

This option systematically samples rotamers for each residue side-chain in the input selection and uses resulting conformational ensembles to evaluate energy-weighted RMSDs for every side-chain atom. These are stored in the 'field' values on atoms and can be used for example to color the structure by side-chain flexibility. Conformational entropy for each residue side-chain is also calculated and stored in a table. If l_entropyBfactor flag is on, the atom rmsds are normalized within the residue to reflect its total conformational entropy. If l_bfactor flag is set, the bfactors are reset to the same values that are placed in the atom 'field', and occupancy is set to be inversely proportional to it ( O=1/(1+2*rmsd) )

• Read pdb file (File/Open or PDB Search Tab).
• Convert to an ICM Object.
• Tools/3D Predict/Local Flexibility

3.6.4 Protein Interface by ODA

The ICM Optimal Docking Area method is a useful way of prediciting likely protein-protein interaction interfaces. If you do not have mutational data or other experimental data which indicates the likely protein-protein docking site this method will be useful. This procedure can save you time during the docking procedure by focusing your docking only on areas on the receptor and ligand most likely to interact.

Theory

ODA (Optimal Docking Areas) is a new method to predict protein-protein interaction sites on protein surfaces. It identifies optimal surface patches with the lowest docking desolvation energy values as calculated by atomic solvation parameters (ASP) derived from octanol/water transfer experiments and adjusted for protein-protein docking. The predictor has been benchmarked on 66 non-homologous unbound structures, and the identified interactions points (top 10 ODA hot-spots) are correctly located in 70% of the cases (80% if we disregard NMR structures). For a description of the method see Fernandez-Recio et al Proteins (2005) 127: 9632.

To display the optimal docking area.

• Convert the PDB file to an ICM object.
• Tools/3D Predict/Protein Interface by ODA
• If you select the Residue Table option the average ODA score for each residue will be displayed in a table. The lower the number the higher the chance the residue will be involved in protein-protein interactions. Regions colored red represent low ODA score and blue represents a high score.

ODA Example with a subtilisin-chymotrypsin complex.

As an example we will determine whether the ICM-ODA method can accurately predict the binding surface of the complex between subtilisin and chymotrypsin. This example is used in the protein-protein docking tutorial below as well.

This complex has been solved experimentally and has PDB id 2sni.

Calculate the ODA for each subunit (Tools/3D Predict / Protein Interface by ODA).

ODA for subtilisin and ODA for chymotrypsin - red colored spheres indicate a region highly likely to be involved in protein-protein interaction, blue coloring is unlikely to be involved in protein-protein interaction. A clickable table is also displayed with ODA values.

3.6.5 icmPocketFinder

If a binding pocket is not known in advance, use icmPocketFinder or icmCavityFinder (for closed pockets). The protein needs to be converted to an ICM object in order to use icmPocketFinder.

icmPocketFinder can be accessed by

• Click on the menu Tools/3D Predict/icmPocketFinder

• Enter a tolerance level (4.6 is the default value and we recommended you to use this). The lower the tolerance value the more pockets predicted and the higher the tolerance the less pockets predicted.
• Check the box create sequence sites if you wish the site to be labeled.
• Check the box display results to see the predicted pockets as grobs in the display panel.
• Check the box keep compounds if you wish the compounds (ligands) in the receptor to be included in the prediction. If you dont check this box the pockets will be calculated based on the receptor without ligands.
• Click OK to run icmPocketFinder

NOTE: A button for icmPocketFinder can be found on the Setup Receptor option in the docking menu. It performs the same function as Tools/3D Predict/icmPocketFinder

The results from icmPocketFinder will be displayed in a table.

To view the pocket in the graphical display:

• Click on the pocket in the table or select the pocket from the meshes section of the ICM workspace. Right click on the pocket mesh in the ICM Workspace to retrieve more display options.

The results from icmPocketFinder are also plotted graphically (Area vs Volume). A blue square highlights potential drug binding pockets based on typical area and volume values - this is only a guide on what constitutes a pocket likely to be involved in ligand binding. Selections can also be made from the plot by clicking and dragging around a point in the graph.

To identify ligand binding pockets which are completely enclosed in the receptor:

• Click on the menu Tools/Analysis/Closed Cavities and a window as shown below will be displayed.

A similar output to that generated by ICMPocketFinder will be displayed. This output includes a plot and a table. By clicking on the table or plot graphical selections can be made.