I have placed several items that you will need for the problem set in the 5.52 c ourse locker.
Draw the complex of trypsin with BPTI (Bovine Pancreatic Trypsin Inhibitor). In this structure a surface loop from BPTI occupies the active site of trypsin. I have already down-loaded this structure from the PDB (entry 2PTC), and split it onto three fi les which you can find in the 5.52 locker: trypsin.pdb (trypsin only, residue numbers 16- 245), BPTI.pdb (BPTI only, residue numbers I1-I58) and trypsin-waters.pdb (H2O atoms f rom both trypsin and BPTI). You may find it easiest to color BPTI a differently fro m trypsin. Examine the active site, using the above figure from "Introduction to Protein St ructure" by Branden and Tooze. Identify the catalytic triad, oxyanion hole, specificity pocket, and non-specific substrate main chain interactions.You may be interested to look also at the structure of trypsin in complex with t he small molecule inhibitor p-amidinophenylpyruvate (PDB entry 1TPP). I have rotated an d translated these coordinates so that they line up with trypsin-BPTI complex, and placed them in trypsin_APA.pdb. These structures were independently solved in different crystal forms, and give an indication of the level of precision in highly-refine d X-ray structure determinations, and also of the level of conformational change seen af ter BPTI binding. Overlay the two structures, and look at the differences between them.
In this exercise we will be mostly concerned with interactions in the specificity pocket. You can find a model peptide (pep-lys.pdb) in the 5.52 home directory, which I derived from BPTI. Draw this peptide, and look carefully at the interactions in the spe cificity pocket. List these interactions.
The "specificity pocket" is actually a depression on the surface of the molecule . Protein surfaces have fractal properties, and are difficult to come to grips with. A construction that I do find useful is called the "molecular surface", or "Connolly surface" ( after its inventor) and is basically what you get by rolling a ball the size of a water mo lecule all over the protein. I have calculated a molecular surface for trypsin in trypsin. srf . You can draw this in Quanta as a locus of dots that define the surface. Draw the su rface.
The steric interactions in the pocket may be easier to see if you draw the lysine side chain from the peptide substrate as a CPK model and draw it with the molecular surface . (You can make a CPK for the whole peptide, but it will move more slowly - a CPK for all of trypsin will move extremely slowy and will be difficult to manipulate)
Redesign the active site of trypsin by mutating residues in the specificity pocket, so as to cleave after phe, tyr, trp (like chymotrypsin), or after val, ala (like elastase ) or after ser, thr. Choose one for your group. (I do not know of a protease with the last sp ecificity - if you choose this one, think about why a protease with this specificity might not have evolved). You should mutate the pep-lys molecule to have an appropriate side chain in the active site to act as a guide for the redesign. Try to design in favorable hydrogen bonds, electrostatic interactions, and van der Waals contacts, while avoiding the creation of cavities and steric clashes. Do not forget the water molecules (you can mov e waters into or out of the site as you wish). You can use the Quanta protein design rotamer, spin, and bump tools, and you can adjust the side-chain torsions manually. Remember to consider the preferred conformations (rotamers) for amino acid side chains, as well as the observed spread around the mean torsion values for each rotamer.
(optional) You can assume that the protein backbone remains unaltered by your substitutions. If you do want to change the backbone, you may want to restrict yourself to conformations that have been observed in other proteins (using the Fragment Data base in the Model Backbone menu). You can also isolate individiual residues or segments by breaking the bonds that connect them to the rest of the proteins, and rotating/translating the fragment. After any manipulation of the backbone, you should regularize the geometry of the new backbone (with Regularize Region).
(optional) You can also minimize the predicted potential energy of your rebuilt molecule if you wish. Comparison of the initial to the minimized coordinates ca n identify regions that have not been built into a favorable (low energy) conformation. The CHARMM package can be used for energy minimization and dynamics, and runs interactively from Quanta, with the results reported to the Quanta textport window and also to a file called CHARMM.LOG. By default the energy calculations will inclu de bond, angle, dihedral and improper angle "internal" geometric terms, and electrostatic and van der Waals "external" non-bonded terms. To calculate the energy for all active ( or displayed) molecules, use Charmm Energy from the Modeling menu. To minimize this energy by slight adjustments of the coordinates, use Charmm Minimization from th e Modeling menu. There are many ways to search conformational space, but the defaults (Adopted-basis Newton Raphson) seem to work well. Probably you will need many minimization cycles to achieve convergence.
Save your coordinates along the way. Quanta will allow you to save the current coordinates by rewriting the old .msf file, by writing a new .msf file with a new name, or by keeping track of several versions of the old file with extensions .001, .002, etc. If you choose this last option be sure to eventually delete all of the uneeded version because they can quickly use up your disk quota.
Check the fit of the specificity pocket in your redesigned protein to model peptides with substituted residues, using surfaces and CPKs or Quanta's neighbor analysis. Rebuild your model protein as necessary.
In class, present the rationale for your redesign. If you have a view that you like, save it with "snapshot" from the unix prompt (place the red box around your picture with the left mouse button, and save it with the right mouse button).