Homology Modelling of Superoxide Dismutase [Cu-Zn]
We are going to build an homology model for the Superoxide Dismutase [Cu-Zn] protein of Drosophila melanogaster (SoD, Uniprot AC: P61851).
The Superoxide Dismutase [Cu-Zn] destroys radicals which are normally produced within the cells and which are toxic to biological systems. It is an homodimer which binds 1 copper ion and 1 zinc ion per subunit. The protein belongs to the Superoxide dismutase (Cu/Zn) family and each chain contains 2 binding sites, one for copper and one for zinc.
Please enter the Swiss-Prot accession code for Superoxide Dismutase (P61851) into the input box of the Modelling menu. The input will automatically get validated:
Lets search now for templates to build a model for the target protein. This is a critical step of homology modelling. Selecting a good template is the main determinant of the model quality. To search for templates please press the “Search for Templates” button. There are several templates (as of February 2014, more than 300) available to build a model for the entire protein:
When the template search finishes, a table with the top ranking templates is presented. As you can see, there are multiple templates which cover the complete sequence and share considerable sequence identity to the Superoxide Dismutase. We are now going to analyse the identified templates in more detail and compare them to each other.
To structurally compare the templates, select for instance the 4 top-ranking templates. They will be superposed and shown in the 3D viewer. What can you say about the structural similarity? To retrieve detailed template information click on the small arrows at the end of the template lines to expand the box with the description of the individual templates. What ligands are present? Is there a template which has all the natural ligands?.
By looking at the alignment (please select the "Alignment of Selected Templates" tab) , you will notice that the sequences of the templates are very similar. To change the coloring, click on the "Options" button (cog icon) and activate the clustal coloring scheme. Pay special attention to insertions/deletions displayed in the target-template alignment:.
To get an even better overview of the sequence identity between the templates, open the "Sequence Similarity" tab. Each template is shown as a circle. The distances between the templates in the plot is proportional to the sequence identity between them. Thus, similar sequences cluster together. Perform a "sweep select" of one of the clusters and analyze the templates structurally. Would you expect large structural variations within a cluster?.
Based on the knowledge you have gathered (insertions/deletions, ligands, oligomeric state), select one or two templates for building a model. Expand the box with the description of the individual templates (click on the small arrow) and check the radio button corresponding to the oligomeric state you want the model to be built. Then click the "Build Model" button:.
Which of the templates would you pick to build a model?
The templates have very similar structures. The largest structural differences can be seen in regions where one of the templates has an insertion and the other does not. These regions locate in loops between two beta strands and thus have little impact on the overall topology of the structures. The majority of the templates have been experimentally solved by X-Ray crystallography. The proteins are from different organisms, some contains mutations. And there are also some differences in the type of ligand they bind. Not all solved structures are bound to both natural ligands Cu and Zn.
We have built a homodimer model for the Superoxide Dismutase based on template 3l9y (chain A), which is the second best hit according to our global quality estimation (GMQE) and contains both Cu and Zn ligands:
Model Output Both global and local estimation of the quality of the obtained model are reasonable: