• Why protein modeling?

• Illustration of modeling method.

Why protein modeling?
 

 
• To understand the function of a protein given only its sequence.

• Sequence => 3D Structure => Function.
 
• To do rational drug design.

• Protein 3D structure dictates which drug can bind to it.
 
 

• Experiments in 3D structure lags far behind sequence data:

• Progress in experimental study of sequences:

• Number of sequences growing rapidly.
 
• 240,000 protein sequences determined so far.
• Complete human genome to be determined by 2003.
 
 
 
 
• Progress in experimental study of 3D structures:

• Structure determination slow
 
• 7,000 structures determined so far.
• New structure rate: ~100-200 per month.
 
 
• Modeling can help:

• Of the 240,000 protein sequences, 70,000 are related to known structures.
 
• Basis:
Protein falls into families:
Similar origins.
Similar functions.
Similar folding patterns.
There are only a total of 1000 protein families.
 
• This relationship may be used to model protein structure.
 
• Strategy:
 
Structure prediction based on template proteins whose sequence is similar to the to be modeled protein.

This strategy is called protein homology modeling.
 

Modeling method:

• Information needed:

 

Protein A with unknown 3D structure:
 


 
 

 
 

Template protein B, C, and D with known 3D structure:
 
 
 
 
  sequence similar to domain 1 in Protein A
 



   sequence similar to domain 2 in Protein A

 
 

   sequence similar to domain 3 in Protein A
 
 
 

• Modeling steps:

Step 1:

Given the sequence of protein A, conduct sequence alignment to find proteins B, C, D, .. in PDB or other commercial databases that are similar in sequence to one or more domains in protein A.
 
 

Step 2:

3D structure of the template for protein A =

Domain i of B +
Domain j of C +
Domain k of D +
Loops connecting these domains
 
 
 
Step 3:

Mutation of the sequence of the template to the exact sequence of  protein A.

Step 4:

Molecular mechanics computation of the mutated template to find low potential energy conformation.

3D strcuture of protein A =

mutated template with lowest potential energy.
 
 
• Available softwares:

• Commercial:
 
• SYBYL.
• Produced by Tripos in the US.
• Our Department has licensed it.
• Insight II.
• Produced by MSI in the US.
• IMCB and NUS computer center may have it.
 
 
• Noncommercial:
 
• Modeller.
• Developed at Rockefeller University in the US.
• Modeller home page and ftp download site.