Since about 1980, the involvement of QM/MM calculations in decoding biochemical processes has been incessantly increasing. On the practical aspect, performing accurate QM/MM calculations on protein systems requires strong dedication and benchmarking involving:
- The test of the method (QM, MM, basis set, force field...)
- The test of the partitioning (where to position the link atoms, what are the best cuts for representing the real system, what are the best cuts to product faster convergences, ...)
Setting up and testing a QM/MM approach on proteinic complex in short practical session would require too much time and effort not only because of the two points mentioned before but also because very few programs (if any) allow a preparation of QM/MM calculations and their analysis in a sufficient user friendly fashion.
In order to give you the possibility to taste most of QM/MM practicabilities and complexities, we will work on the particular problem of the reaction of the chromophore of the antiobiotic antitumor protein: Neocarzinostatin(Kim et al.). Prior to the study have also a look at the work of Myers et al. in J. Am. Chem. Soc. 1996 and discuss together what would be the way to understand the results of this work using QM/MM approach.
- Go to the Protein Data Bank and search for a holo form of the NCS (pdb 1O5P or 1nco). Open the holo-NCS pdb in Chimera. Look at the general forl of the protein and identify the structure of the chromophore. Discuss briefly and elaborate molecular hipothesis with respect of Myers' paper (possible sites of interaction of the thiol, paricipation of water molecules, movement in the binding sites (compared both structures to do so) etc...).
- Imagine, our final objective would be to perform the study of the catalytic mechanism in the NCS with thiol when the chromophore is in the binding site of NCS. What would be the number of atoms you will need to perform a reasonable representation of the system protein-chromophore and protein-chromophore-thiol?
4. Since the the chromophore on its own is substantially large (81 atoms heavy + hydrogen), one would like to get the possibility to already partition the chromophore. That will be our task from now on. From the pdb file save only the chromophore (from pymol or editing the pdb file)
Open Gaussview and check that the structure of the chromophore has the correct bond order and the correct atom connection. Gaussview gives many problems in the preparation of the input so check if your gaussian input has the correct atom types for the link atom and the correct connection. The tutor will carefully explain you the main traps of the inputs.
5. Our first calculation is to minimize the structure of the chromophore in a am1:dreiding environment using oniom. Prepare the input with the help of the tutor and acoording to the partitioning you had in mind. Share your calculations with you classmates in order to have a vast panel of partition to discuss. Optimize your structure and look at the following:
1. does it converge?
2. what happened?
6. Getting further.
1. Do comparative analysis with your classmates
2. extend your work to the Thioadduct if possible
3. set over night the calculations with an abinition method in the electronic part
4. Try a QM/QM/MM approach
7. Do your own experiments!