BIOL 701 Online Homology Modeling Exercise

Exercise
  1. In this exercise, you will be generating and analyzing a homology model of the gamma tubulin protein, with the following sequence:
    MPSEIITLQLGQCGNQIGFEFWKRLCLEHGISPSGVLEDFANDGLDRKDVFFYQADDDHY
    IPRAVLLDLEPRVINTIMGSVYSKLYNPENVYLSKHGGGAGNNWASGYSQGEKLQEEVFD
    IIDREADGSDSLEGFILCHSIAGGTGSGMGSFIMERLADRYPKKLIQTFSVFPNQDEISD
    VVVQPYNSMLTLKRLTTAADSVVVLDNTALNRIACDRLHIQNPSFSQINNLVSTIMSVST
    TTLRYPSYMNNNLIGLTAPLIPTPQLHFLMTGYTPLTSDSDIHTQQLVNVRKTTVLDVMR
    RLLQPKNMMVSTGPDKSNHHCYISILNIIQGEVDPTQVHKSLQRIRDRKMAQFIPWGPTS
    IQVALSRSSPYVQSNHRVSGLMLANHTSICSLFERALNQYDKLRKRGAFLDQFRREDIFK
    DDLNELDESRETVDCLVQEYEAATREDYMQFSVKRGNGPVDSKSEDSRSVTSAGS
    
  2. In another browser window, go to the following URL:
    http://swissmodel.expasy.org/

     
  3. Click on the link (left side bar) called "First Approach mode":
    (http://swissmodel.expasy.org/SM_FIRST.html)

     
  4. Fill in the following fields in the form:
    • email address
    • your name
    • a title for your modeling job
    • paste the sequence from step 1)
       
  5. Scroll down the form until you see the "Results Options:" list and:
    • select "Normal mode"
    • select "Include a WhatCheck report"
       
  6. Scroll back up the form and hit the "Send Request" button. 
     
  7. The results of your automated homology modeling will be sent to you via e-mail (may take up to a half hour or so). You will receive a "TraceLog", a "Model" and a "WhatCheck". 
     
  8. Download the structure file given as an attachment in your Model output. Take note of the directory/folder where you save it.
     
  9. Upload your *.pdb structure file from step 8) into the applet. To do this, either type in the full path name for your file, or else use the "Browse..." button. Note that when browsing, you may need to change the file filter to "*.pdb" since most browsers are looking for *.htm or *.html files by default. 
     
  10. Take a look at the model you have generated. Familiarize yourself with the way the different mouse buttons work to move the model around. 
     
  11. Using the "Select" button on the right of the widget, bring up the "Select" widget. 
     
  12. The Select Widget comes up as a separate window, but is still capable of interacting with the original view of the structure. It provides many useful features including the ability to see where individual residues are located. In order to highlight a specific residue in your structure:
    • using your mouse, highlight the residue of interest on your Select Widget
    • change Method to "SC vs. BB" (side chain vs. backbone)
    • change the color to something not already in use (e.g., lilac)
    • press "Apply"
    The representation you see has the full amino acid of interest rendered in the special color, and all other amino acids in green, showing only their backbone. Once you've figured out where your amino acid of interest is located, you can turn the full amino acid representation back on by changing "SC vs. BB" to "SC vs. SC". The residue of interest should still be in the new color, but the rest of the protein will appear in full detail.

Questions

  1. Review the material in your trace log output and answer the following questions:
    1. How many templates were used to generate your structure? How good were their respective sequence identities relative to the gamma tubulin model? Comment briefly on the implications of the number of templates and the relative identity scores on the quality of the alignment.
    2. What is the total energy of the model? Remember that negative energies represent conformations where favorable electrostatic and van der Waals interactions outweigh unfavorable interactions, while positive energies represent cases where strained interactions outweigh attractions. What does the energy of your model tell you about the stability of the predicted structure?
  2. Review the material in your WhatCheck output and answer the following questions:
    1. Do you see any cases of unnatural bond lengths, angles or torsions? What sort of molecular modeling technique could you use to refine the predicted structure so as to get rid of such abnormalities?
    2. Hydrogen bond donors and acceptor on the surface of proteins are often unsatisfied (i.e., a donor does not have a complementary acceptor or vice versa). This is normal, since H-bond donors and acceptors are convenient sites by which proteins can interact with solvents or other proteins. Subsurface (i.e., buried) side chain H-bond donors and acceptors are almost always satisfied in proteins however, since this is part of what stabilizes the protein. Are there any unsatisfied side chain H-bond donors or acceptors in your system? If so, what does this tell you about the quality of the model?
    3. Using WebMol (steps 9-13 above), take a look at the residue "2 E" (i.e., the glutamate in sequence position 2) which is probably one of the ones you would have found to have an unsatisfied H-bond in question 2b). Based on the position of this side chain, indicate whether you think this side chain is unsatisfied because a) it probably should be on the surface but the model mis-oriented the side chain, or b) it really should be subsurface, but the model mis-oriented the side chain of its H-bond partner. (Hint: is the residue close to the surface? Or does it have a nearby unsatisfied base that is should be partnered with?)

CONTACT
David Johnson
emaildkjohnson@ku.edu
Access to more than 40 computational chemistry software programs and databases
High-performance computational tools accelerate drug discovery and minimize costs
Analyze complex, multidimensional data sets to rapidly generate biological insights
KU Today