Tutorial

More tutorials

Binding Site Comparison

CONTENT

 1.  Goal

 2.  Introduction

 3.  Input

 4.  Chain selection

 5.  Box selection

 6.  Probe and algorithm selection

 7.  Runtime estimation

 8.  Output visualization

 9.  Output interpretation

10.  Reference

1.  GOAL

Go to top

- To identify regions wherein two proteins exhibit difference in ligand-binding properties;
- To gain information to guide protein engineering and drug design for higher binding selectivity.

2.  INTRODUCTION

Go to top

  SiteComp conducts point-to-point comparison of the molecular interaction fields (MIFs) of two superposed binding sites and highlights the differences in molecular interaction properties. Any difference detected by SiteComp lies in the observed difference in MIFs, which reflects both sequence and conformational changes. Moreover, to ensure meaningful results, SiteComp should only be used to compare structurally similar binding sites.

  In this tutorial, SiteComp is used to compare the binding site MIFs of Cyclooxygenase (COX) enzymes. SiteComp identifies a difference cluster which is confirmed to contribute to binding selectivity. COX enzymes are targets for inhibition by non-steroidal anti-inflammatory drugs. Selective inhibition of COX-2 is wanted in treating patients susceptible to complications related to COX-1 inhibition [1]. COX-2 (PDB: 5cox, Mus musculus) and COX-1 (PDB: 1eqg, Ovis aries) are superposed and their binding sites are compared.

  The example output page used in this tutorial can be found here.

Note: If both bound and unbound structures are available, like in this example, it is recommended to use structures with the same or similar ligands in both proteins. Otherwise, it is preferable to use the unbound structure of the protein for which selectivity is desired. In this example, the COX-2 structure is unbound, and the COX-1 structure is bound to a non-selective inhibitor, Ibuprofen (IBP).

3.  INPUT

Go to top

On the input page:

(1) Set task as "compare two proteins ".

(2) Enter two protein PDB IDs or upload two PDB files. For example, enter "5cox " and "1eqg ".

(3) To use structures superposed by other programs, select "The uploaded structures are pre-superposed " to skip superposition later.

(4) Click on "Submit ".

4.  CHAIN SELECTION

Go to top

(1) If there are more than 1 chain in either protein, select chains for calculation. Unselected chains will be excluded for calculation. In this example, select the A chains from each protein.

(2) Click on "Submit ".

Note: This step is skipped if neither protein contains more than 1 chain.

5.  BOX SELECTION

Go to top

Define a box-shaped region within which comparison will be performed. Both proteins should participate in the box with some equivalent residue pairs. To ensure meaningful results, use a relatively small box enclosing the region of interest. In this example, this should be the inhibitor binding site. It is recommended to leave a margin of a few angstroms in each direction.

(1) For better visualization, use simple mouse actions and buttons in Section 1 to manipulate the molecules in Jmol applet.

(2) Box center could be set in Section 2 in three ways: 1. clicking on an atom in Jmol applet; 2. specifying a residue number; 3. specifying center coordinates. The center of the box will be moved to the specified atom/residue/coordinates. The default box center is the geometric center of the protein.

(3) Box dimensions could be set in Section 3 (Unit: angstrom). The default box dimensions are 0.3 times the dimensions of the protein.

Initial View:

Large image

Manipulated View (obtained with the following steps: 1. hiding protein backbones, showing and labeling heteroatoms; 2. moving box center to <25.100, 22.000, 16.700> by clicking on IBP in Jmol applet; 3. resizing box to 16*16*16 to encompass IBP; 4. rotating the molecule by dragging in Jmol applet):

Large image

(4) It is recommended to refine superposition based on atoms within the box - click on "Refine " in Section 4.

(5) As soon as the box is chosen, click on " Continue " in Section 4.

6.  PROBE AND ALGORITHM SELECTION

Go to top

(1) Select a probe for EasyMIFs calculation. For example, select "CMET" to detect any difference in MIFs regarding methyl-like ligands.

(2) Select a clustering algorithm. For example, select "Average Linkage".

(3) To go back and modify the box, click on " << ". Otherwise, click on " >> ".

Note: Please refer to EASYMIFs & SITEHOUND User's Guide for details about the different probes and clustering algoritms.

7.  RUNTIME ESTIMATION

Go to top

Upon submission, SiteComp provides a runtime estimation and a link to the output page. You will be redirected to the output page when calculation finishes (usually within a few minutes). The output page will be kept on the server for 30 days, during which period it can be revisited as many times as desired.

8.  OUTPUT VISUALIZATION

Go to top

(1) For better visualization, use simple mouse actions and Section 2 to manipulate the molecules in Jmol applet. The size of Jmol applet could also be altered.

(2) Use Section 1 to visualize each set of difference clusters. Available options include:
  - visualize and zoom in to each difference cluster;
  - visualize and label protein sidechains surrounding each difference cluster;
  - visualize and label equivalent residues in the less favorable protein;
  - show surfaces of difference clusters;
  - color difference clusters by relative or absolute energy.
If colored by relative energy, the points with lightest and darkest colors are the ones with least and most favorable energy values within each difference cluster. If colored by absolute energy, the points with lightest and darkest colors are the ones with least and most favorable energy values from each set of difference clusters.

(3) Information of difference clusters, including energy, energy range, volume, center coordinates and participating residues, is shown in Section 4. Residue information could be shown or hidden using the "Show/Hide residue information " button.

(4) By default, the information about the clusters favorable for the wildtype struture is displayed. If there are any difference clusters favorable for the mutant structure, use the "Show details " buttons in Section 2 to display the relevant information. This button is hidden if there is no such difference cluster.

Initial View:

Large image

(5) Use the "DOWNLOAD " link in Section 2 to download output files.

* Explanation of files in the downloaded package   

Note: The .pdb files can be used to display the proteins, heteroatoms and difference clusters in molecular visualization softwares such as PyMol or Jmol. Please refer to EASYMIFs & SITEHOUND User's Guide for detailed description of each file type.

(6) If you wish to go back to the previous step, click on " << ".

(7) To study each difference cluster in detail, specify a difference cluster in Section 4 and click on " Go ". Then follow the tutorials for binding site decomposition or multi-probe characterization.

9.  OUTPUT INTERPRETATION

Go to top

  In Manipulated View 1, the 1st difference cluster favorable for COX-2 (5cox) is displayed. MIF comparison shows that this region in COX-2, compared to the corresponding region in COX-1 (1eqg), is more favorable for methyl-like ligands. Three mutations between COX-2 and COX-1 (R513H, A516S, V523I), as well as displacement of other sidechains, contribute to this difference cluster.

Manipulated View 1 (obtained with the following steps: 1. hiding protein backbones, showing and labeling heteroatoms; 2. zooming in to the first difference cluster favorable for 5cox; 3. showing the surface of difference clusters; 4. showing sidechains for this difference cluster; 5. showing and labeling equivalent sidechains in 1eqg; 6. hiding other difference clusters, hiding boundbox; 7. rotating the molecule by dragging in Jmol applet):

Large image

  This led to the hypothesis that this difference cluster, especially the points with largest differences (shown in darkest colors in Manipulated View 2), may be the key to design a COX-2 selective inhibitor. Since this region is in the vicinity of the non-selective inhibitor, Ibuprofen, one could try to extend Ibuprofen into this region to improve COX-2 selectivity.

Manipulated View 2 (obtained with the following steps: 1. hiding protein backbones, showing and labeling heteroatoms; 2. zooming in to the first difference cluster favorable for 1pwm; 3. coloring this difference cluster by relative energy; 4. hiding other difference clusters, hiding boundbox; 5. rotating the molecule by dragging in Jmol applet):

Large image

  Indeed, this hypothesis was proven by the existence of a COX-2 selective drug marketed by Pfizer, Celecoxib [2]. The structure for COX-2 bound to Celecoxib (PDB: 3ln1) was superposed to the structure of COX-1 bound to Ibuprofen (PDB: 1eqg). A closer look at the binding sites shows that Celecoxib not only occupies the Ibuprofen subsite, but also extends into the region of the first difference cluster identified by SiteComp.

The Celecoxib binding site:

Large image

  To analyze this difference cluster in further detail, one may specify this difference cluster in Section 4 and click on "Go ". Then follow the tutorials for Binding site decomposition or Multi-probe characterization.

10.  REFERENCE

Go to top

1

Grosser, T. (2006) The pharmacology of selective inhibition of COX-2, Thromb Haemost, 96, 393-400.

2

Penning, T.D., Talley, J.J., Bertenshaw, S.R., Carter, J.S., Collins, P.W., Docter, S., Graneto, M.J., Lee, L.F., Malecha, J.W., Miyashiro, J.M., Rogers, R.S., Rogier, D.J., Yu, S.S., AndersonGd, Burton, E.G., Cogburn, J.N., Gregory, S.A., Koboldt, C.M., Perkins, W.E., Seibert, K., Veenhuizen, A.W., Zhang, Y.Y. and Isakson, P.C. (1997) Synthesis and biological evaluation of the 1,5-diarylpyrazole class of cyclooxygenase-2 inhibitors: identification of 4-[5-(4-methylphenyl)-3-(trifluoromethyl)-1H-pyrazol-1-yl]benze nesulfonamide (SC-58635, celecoxib), J Med Chem, 40, 1347-1365.

For help and questions, please email sitecomp@sanchezlab.org