Based on previous exercice developed by Rein Aasland, Morten Mattingsdal, and Pål Puntervoll
Last updated: NOV-2015 by Angele ABBOUD
In this exercise you will continue to work on the cyclin protein family. The focus will be on the protein structure in general and structural aspects of the cyclins in particular.
- 1 Secondary Structure Prediction
- 2 Visualizing structures with Jmol
- 3 Home work: Comparing structures
Secondary Structure Prediction
As we have seen in the previous PC lab, in addition to being a multiple sequence alignment editor, Jalview can function as a client to Web Services offering multiple sequence alignment (e.g. Muscle and ClustalW) and secondary structure prediction: JNet.
Some secondary structure prediction methods use multiple sequence alignments as basis for prediction. JNet (trough Jalview) can be used in two ways:
- Submit an alignment. If you submit an alignment, this will be used directly by JNet for secondary structure prediction, (and the PSI-Blast step will be skipped).
- Submit a single sequence. If you submit a single sequence, JNet will search for similar sequences (using PSI-Blast), create a multiple sequence alignment, and perform the secondary structure prediction based on the alignment. The alignment with secondary structure predictions will be returned to Jalview.
Submit an alignment
Open the alignment of the human cyclins that you obtained at the previous session using the alignment program ClustalW with a penalty gap of 20 - in Jalview Multiple Sequence Alignments
Perform the secondary structure prediction for CCNA2_HUMAN using your alignment:
- Select CCNA2_HUMAN and with the up-arrow of the keyboard, move it to the top of your alignment
- Select all sequences
- Start a JNet job: WEB SERVICE>SECONDARY STRUCTURE PREDICTION>JNET...
- How many of alpha-helices are predicted (judging from the jnetpred result)?
- How many alpha-helices have been predicted but are not represented in the final jnetpred result?
To assess the results of the secondary structure prediction, display the secondary structure features for the sequences, using the Uniprot database as we have done it in the previous session Multiple Sequence Alignments.
- On the alpha-helices predicted in the jnetpred result, how many of them overlap with actual alpha-helices displayed in the annotation?
Submit a single sequence
Perform the secondary structure prediction for CCNA2_HUMAN using a single sequence:
- Open the file with all the cyclin human sequences called cyclin_blast_human.fasta
- Select CCNA2_HUMAN
- Start a JNet job: WEB SERVICE>SECONDARY STRUCTURE PREDICTION>JNET...
- Answer the question Q1 and Q2 regarding this secondary structure prediction now.
- Why does one of the approaches seem to provide much better results than the other? [Hint: What sequences are included in the ccna2_human_jnet.jar alignment?]
Visualizing structures with Jmol
In this part of the exercise we will use the structure viewer Jmol to get acquainted with the three dimensional structure of cyclin. We will look at a cyclin A (CCNA2_HUMAN) in complex with cyclin-dependent kinase-2 (CDK2_HUMAN) and the CDK2 inhibitor P27 (CDN1B_HUMAN). The PDB identifier is 1JSU.
Start by going to the PDBsum web site to get acquainted with the available information on this structure. To find the structure, you can search with the identifier. Download the PDB-file by going to the RCSB/PDB web site.
Spend some time familiarizing yourself with the Jmol resources:
- Jmol main web page.
- Documentation can be found on the Jmol Wiki pages and you might find this one-page Jmol Quick reference useful.
- You will frequently need the Jmol reference. This is a hyperlinked help file which contains explanations for all commands in Jmol and how to select atoms and residues and predefined sets such as hydrophobic residues and helix.
- The RasMol Quick Reference Card is also very useful. Most commands, if not all, will also work in Jmol.
Start Jmol and open your PDB-file of the molecule 1JSU: FILE>OPEN. Note that the Jmol menu for manipulating the structure is available by right-clicking in the structure window. Use the mouse to rotate the molecule. Find the mouse and key combinations for rotation around X, Y and Z axes, and how to zoom and move (translate) the molecule in the window.
Hint: Keep a window with the PDBsum information open for 1JSU to help navigate.
Identify the chains in the structure
First of all, open the Jmol consol window: from the Jmol menu (right-click with the mouse in the structure window) choose CONSOL. The consol or command window is where you communicate with Jmol by issuing commands. In the following, Jmol commands are written with
typewriter font. Comments to commands are written in in parenthesis and should not be typed.
The three chains (proteins) in the 1JSU structure are called A, B and C.
Give each chain a different color. Example:
> select *A > color blue
- Which chain corresponds to which protein?
Valid colors can be found in the jmol documentation.
Note that the last selection you make is always active until you say select all or select something else. Issue the following commands, and observe the effects noted in parenthesis:
> select all > cartoon off; cpk off (to remove the details and start from scratch) > wireframe > select *B (chain B) > wireframe off > cartoon (to show the cartoon representation of the structure) > select helix and *A (to select all alpha-helices in chain A) > color orange > select sheet and *A (to select all beta-strands in chain A) > color violet (Note that the kinase has two subdomains, one alpha-helical and one beta-sheet.) > select *C (chain C) > cpk (space filling representation) > color cpk (standard coloring)
Jmol remembers all your commands, and you can recall them and correct them using the ARROW UP and ARROW DOWN keys. Also, at any time, you can save all your commands to a file which you can load later one to resume where you left (save is:
write script <file name>. load is:
script <file name>).
Identify the hydrophobic core in the cyclin
Issue the following commands:
> select *B > color cpk > cpk > select hydrophobic and *B > color yellow > slab on
The last command activates slab mode. While keeping control+shift keys pressed, use the mouse (hold left button) to peek in and out in the structure. Note that there are large areas in the structure that appear compact yellow. Click on some of the yellow and non-yellow residues and note, in the command window, which residues they are.
Turn slab mode off (
> slab off) and select chain C and render it in spacefilling. Give it a different color from chain B and note how closely the two chains are packed. Do the same for chain A. Now, turn on slab mode again, and inspect the interaction surfaces between the three chains.
- Where are located these compact yellow regions? Which amino acids composed these regions?
Remove the structure using the zap command and load the structure again (as above).
Inspect the functional site motif RxL in the cyclin inhibitor, P27
The P27 cyclin inhibitor (CDN1B_HUMAN) harbours an RxL motif (ELM:DOC_CYCLIN_1), that is used in the interaction with cyclin. Use the ELM server to identify the DOC_CYCLIN_1 motif in P27 (UniProt:CDN1B_HUMAN). Read more about this motif by clicking on the DOC_CYCLIN_1 name in the result page.
- What residues in P27 constitute the DOC_CYCLIN_1 motif?
To reveal the motif side chains in the structure, return to Jmol and do:
> select all > cartoon off; wireframe off; cpk off > backbone 120 > color chain > select 30-33C > wireframe 80
Zoom in and identify each of the four residues in the motif. Note how they project into the structure. Render the cyclin (chain B) in spacefilling and inspect the structure to find what kind of residues in the cyclin form the binding pocket for the RxL motif. It is useful to
> set ambient 24
to highlight the surface topology. Since the RxL motif has a basic residue, we may expect that the binding pocket contains some acidic residues. To highlight the acidic residues, use:
> select acidic > color red
To help identifying the which acidic residues interact with the arginine 30, render it in spacefilling and give it a different color:
> select arg30C > cpk > color cpk
Click on each of the atoms to identify the nitrogen atoms of the arginine side chain.
- Which are the acidic residues in the pocket that recognizes the arginine side chain in the RxL motif?
Now, use this information and the multiple alignment you made using ClustalW (section Structure_Analysis#Secondary Structure Prediction Secondary Structure Prediction) to answer the following question:
- Which of the cyclins are likely to interact with the P27 inhibitor through the RxL motif?
The aromatic residue in the RxL motif is also important:
> select phe33C > cpk
Inspect the region around this Phenylalanine and try to identify critical residues for recognition of this Phenylalanine.
- Are any of these residues conserved in the alignment?
Home work: Comparing structures
The 3D structure is resolved for two of the human: CCNA2_HUMAN and CCNE1_HUMAN. Their PDB code are 1JSU and 1W98, respectively.
Note: Both structures are complexes of the cyclins and other polypeptide chains. The cyclins are in both cases chain B. Hence, you should only use the B chain for the structural comparison below.
To explore structural similarity, perform a pairwise structural comparison using the DaliLite server:
- Enter the PDB codes for the two structures. Remember to also enter the chain code.
- What is the root mean square deviation (RMSD) for the best structural alignment of the two structures?
* Inspect the resulting structural alignment, and compare it to your Clustal alignment (section 1).
- Assuming that the structural alignment generated by DaliLite is the correct alignment, can you identify which parts of the Clustal alignment that are likely to be erroneous?
Is the assumption in Q11 reasonable?