MODPROPEP is a web server for knowledge-based modeling of protein-peptide complexes, specifically peptides in complex with MHC proteins and kinases. The available crystal structures of protein-peptide complexes in PDB are used as templates for modeling peptides of desired sequence in the substrate binding pocket of MHCs or protein kinases. The substrate peptides are modeled using the same backbone conformation as in the template and the side chain conformations are obtained by the program SCWRL.
MODPROPEP can be used to perform the following operations involving the modeling of protein-peptide complex.
1. Modeling of a peptide(s) of desired sequence in complex with MHCs or protein kinases.
2. Scanning a protein sequence for potential binder peptides of MHCs or protein kinases using a scoring scheme involving residue based statistical potential. Selected peptides from the protein can be modeled in complex with an MHC or a protein kinase of choice.
3. Transfer of the coordinates of a crystal structure bound peptide from a homologous protein to a structure without the bound peptide ( in case no crystal structure in complex with a peptide is available). The resulting complex can be used as a template for modeling the sequence of desired peptide, or scanning a protein sequence for potential binder.
4. In case no crystal structure is available for an MHC or protein kinase (list of structural templates), MODPROPEP can build a homology model using a homologous structure from the MODPROPEP structure library.
5. MODPROPEP allows the analysis of the modeled complex by identifying detailed inter molecular interactions between the peptide and the protein.
6. It allows the visualization of protein-peptide complexes using a very user friendly Jmol Java Applet.
Please refer following publications for detailed discussion on development and benchmarking results for prediction of phosphorylation sites
Development of MODPROPEP:
Narendra Kumar and Debasisa Mohanty (2007) MODPROPEP: a program for knowledge based modeling of protein-peptide complexes. Nucleic Acids Res, 35: W549-555. PDF PUBMED
Benchmarking results for prediction of phosphorylation sites:
Narendra Kumar and Debasisa Mohanty (2009) Identification of substrates for Ser/Thr kinases using residue based statistical pair potentials. Bioinformatics, 10.1093/bioinformatics/btp633.
The following figure shows the flowchart depicting the organization and features of MODPROPEP.
Currently MODPROPEP can model the protein-peptide complexes involving MHCs and protein kinases only. The program contains three major sections for modeling of complexes involving MHC Class-I, MHC Class-II and Protein Kinases respectively. They can be accessed by clicking the links on the horizontal bar just below the header graphics. The fourth section contains this tutorial. This tutorial describes the use of MODPROPEP for carrying out various task involving protein kinases. It can be used in the similar way for carrying out various tasks involving MHC Class-I and MHC Class-II.
1. Clicking on the “Home” takes you to the home page of the program.
2. Click on "MHC Class-I" for modeling protein-peptide complexes involving MHC Class-I.
3. Click on "MHC Class-II" for modeling protein-peptide complexes involving MHC Class-II.
4. Click on "Protein Kinase" for modeling protein-peptide complexes involving Protein Kinases.
5. Click on "Tutorial" for this tutorial.
1. Select a protein kinase from the list of available protein kinases for which crystal structure in complex with a peptide is available.
2. Select the length of the peptide you want to model (this is seven for protein kinases with three amino acid on the each side of phosphorylation site).
3. Select a task to perform as “MODEL PEPTIDE”.
4.
Click
to go to next page to choose a crystal structure template from available
structures of protein kinase chosen in step1.
5. Select a template from the available crystal structures as template for building protein-peptide complex.
6. Enter the sequence of peptide you want to model in single letter code. If you want to model more than one peptide, please give one peptide sequence per line.
7.
Click
to model the peptide in complex with the chosen crystal structure template.
Peptide will be modeled using the same backbone conformation as in the template
complex.
9. Choose a distance criteria for assuming two residues between peptide and protein kinase to be in contact. Two options are available – first, “ANY CLOSEST ATOMS” in which case two residues are assumed to be contacting if any two atoms between are at a distance less than the cutoff value, and second, “C-BETA” in which case the two residues are assumed to be interacting if the distance between their cβ atoms is less than the given cutoff distance.
10. Choose the cutoff distance in angstroms.
11. The default scoring scheme for evaluating the interaction between peptide and protein kinase is Miyazawa and Jernigan Pair Potential Matrix. Betancourt and Thirumalai scoring matrix is also available as an alternate scoring matrix. Optionally you can upload your own scoring matrix for evaluation of the interaction. The matrix should be uploaded in a predefined format. An example file of scoring matrix is given here. The uploaded matrix file should be exactly in the same format.
12.
Click
to analyze the protein-peptide complex.
13.
This page
lists all the neighboring residues in protein kinase for each residue of modeled peptide
and compares it with the neighboring residues in the template structure.
Clicking on
residue
name will cause the visualization of the residue and the neighboring residues
in the Jmol Java Applet. User has the option to rotate, zoom and other visualization
option provided by the Jmol Applet.
14.
Clicking on
residue
no. opens a pop-up window displaying a list of all atomic contacts between the
clicked residue of the peptide and all its contacting residues in the protein kinase.
1. Select a protein kinase from the list of available protein kinases for which crystal structure in complex with a peptide is available.
2. Select the length of the peptide you want to scan from the amino acid sequence of the substrate protein (this is seven for protein kinases with three amino acid on the each side of phosphorylation site). The entered protein sequence will be broken into all possible 7mer peptides containing phosphorylation site at 4th position of the peptide.
3. Select the task to perform as “RANK PEPTIDES”.
4.
Click
to go to next page to choose a crystal structure template from available
structures for protein kinase chosen in step1.
5. Select a kinase structure from the available crystal structures as template for scanning the substrate protein sequence for potential phosphorylation sites.
6. Choose a distance criteria for assuming two residues between peptide and kinase to be in contact. Two options are available – first, “ANY CLOSEST ATOMS” in which case two residues are assumed to be contacting if any two atoms between are at a distance less than the cutoff value, and second, “C-BETA” in which case the two residues are assume to be interacting if the distance between their cβ atoms is less than the given cutoff distance.
7. Choose the cutoff distance in angstroms.
8. Paste the substrate protein sequence to be scanned for potential phosphorylation sites.
9. The default scoring scheme for evaluating the interaction between peptide and protein kinase is Miyazawa and Jernigan Pair Potential Matrix. Betancourt and Thirumalai scoring matrix is also available as an alternate scoring matrix. Optionally you can upload your own scoring matrix for evaluation of the interaction. The matrix should be uploaded in a predefined format. An example file of scoring matrix is given here. The uploaded matrix file should be exactly in the same format.
10.
Click
to score all possible phosphorylation sites in the pasted substrate sequence.
11.
Select the
peptides you want to model in complex with the selected protein kinase and click
to model these peptides. The next page shows the seleceted peptides.
12.
Click
to model selected peptides.
The rest of the steps are same as the step 8 onwards under section “Model the peptide of desired sequence in complex with a protein kinase”.
If no crystal structure is available in complex with a substrate peptide, a peptide can be modeled by transferring the ligand peptide from a homologous structure in the MODPROPEP structure library. The peptide is transferred after optimum superimposition of these two structures. The resulting complex is used as a template for modeling protein-peptide complexes and scanning substrate protein. In order to test the accuracy of this ligand transformation approach, we modeled a peptide in complex with PKB by transforming the bound peptide from PKA. The figure below shows the modeling of a peptide in complex with PKB (template:1O6K) by transforming the bound peptide from the crystal structure of PKA-peptide complex (template:1JBP). Modeled peptide is shown in magenta while original PKA bound peptide is shown in blue for comparison. PKB is shown in cartoon representation in green.
The following steps describe how to carry out the transfer of ligand peptide in MODPROPEP.
1. Click
to go to the peptide transfer page.
2. Select a protein kinase without a bound peptide.
3. Select a structural template for protein kinase selected in step 2.
4.
Click
to choose a homologous protein kinase structure with peptide from which peptide
is to be transferred.
This page displays the protein kinase with bound peptide from structure library of MODPROPEP which is most similar to the protein kinase chosen (without peptide). These two structures are superimposed on each other and the coordinates of the peptide are transferred to the protein kinase without the bound peptide. Alternatively you can choose your own structure from which you want to transfer the peptide.
5. Click “GET SEQUENCE FILE” to get the sequence of protein kinase chosen.
6. Click "View Blast Alignment" to get the blast alignment of "protein kinase without peptide" with "protein kinase with peptide".
7. Click “TRANSFER PEPTIDE TO XXXX FROM YYYY” button to transfer peptide to XXXX structure from YYYY structure.
8.
Or choose a
different kinase from which you want to transfer the peptide. Then click
button to transfer the peptide from this structure.
9. Click "Download Model of Kinase – Transferred Peptide Complex” to download the coordinates of protein kinase—transferred peptide complex in PDB format.
10.
Click
to view the complex in Jmol Java Applet.
11. This model is used as a template for carrying out further operations such as modeling peptide of desired sequence or scanning the substrate protein.
If no crystal structure of the protein kinase of interest is available, MODPROPEP can build its structural model using a homologous protein kinase as template. The homologous structure is identified on the basis of sequence similarity with the query kinase. The following steps describe how to model a protein kinase using MODPROPEP.
1.
Click
button to access the kinase modeling page.
2.
Paste the
amino acid sequence of the protein kinase to be modeled in the text box. Click
button to model the kinase. The next page gives information and links related to
the modeled kinase.
3.
Click
for obtaining the sequence file of the entered protein.
4.
Click
to view the blast alignment between the entered protein sequence and template
used to build the homology model.
5. This model is used as a template for performing further operations such as modeling peptide of desired sequence or scanning the substrate protein.