In this tutorial, students will progress in the most logical fashion from learning about 1) the properties of the small molecule substrate that binds to an enzyme to 2) the intermolecular interactions between the substrate and the enzyme to 3) the amino acid residues and cofactor molecule that directly catalyze the reaction. All of this will set the stage for an investigation into the basis for stereoselectivity in the (R)- and (S)-Hydroxypropylthioethanesulfonate Dehydrogenases (HPCDH). The use of comparative protein structure modeling (step 5) is essential since the structure is only known for R-HPCDH. Each of the steps includes background material to help clarify important concepts.
The structure of R-HPCDH has been determined by x-ray crystallography and deposited into the Protein Databank (PDB) with a code 2CFC. We often refer to the text file that contains the coordinates a PDB file and these files have a particular format that is read by viewers and other programs. The crystal contained the tetrameric enzyme, the bound nicotinamide adenine dinucleotide (NADH) cofactor, water molecules and the product of catalysis, 2-ketopropylthioethanesulfonate (KPC). The structure is very informative for understanding how the enzyme catalyzes the oxidation of hydroxypropylthioethanesulfonate (HPC) but requires a little more experience to distill this information than is typical for students new to protein structure and enzyme mechanisms. For one thing, protons attached to the carbon, nitrogen, oxygen and sulfur atoms are not visible in electron density maps obtained from x-ray crystallography and therefore not normally included in any deposited PDB structures. This omission makes it difficult to demonstrate intermolecular interactions (hydrogen bonding!) and the enzyme mechanism (a proton and hydride transfer).
Therefore, we have used various molecular modeling programs (CHARMM, MMTSB) to construct “all-atom” models of the enzyme-substrate system based on the crystal structure. We then performed molecular dynamic (MD) simulations to produce a model that we believe recapitulates the Michaelis complex (a reactive configuration). A tutorial on the construction of these systems for MD simulations is in the works for students of computational biophysics but is probably not relevant to a course in sequence-based bioinformatics or possibly even structural bioinformatics.
Finally, although we show an example of how to use the program MODELLER to construct a comparative protein structure model, the applicability of this example for other comparative protein modeling projects obviously depends on the details of the project. Computational structural biology software can be difficult for students new to this field (especially if you don’t even bother to read the manual!). The main problem is often not the program but the particulars of the PDB file given to the program as input. Often the PDB file will have some or all of the following complications:
In general these problems can be overcome by manipulating the sequence or structure text files but can require some determination and patience. Students can also choose to submit their target sequences to many good web-servers that will select appropriate template structures, align the template and target sequences and return in cases of strong sequence similarity good 3D models. Instructors may wish to use these servers (links given at the end) for students attempting to model their own sequence instead of using MODELLER as we do in this tutorial.