There are several steps in any enzyme-catalyzed reaction; these are minima on the potential energy surface, and, except for the first and last steps, are usually referred to as intermediates. There are about six steps in the mechanism for the hydrolysis of a peptide bond using chymotrypsin, a well-understood process. In other enzyme-catalyzed reactions the mechanism might be less well understood. This is, however, not a serious limitation, in that an attempt can be made to construct a mechanism, and that mechanism can then be modeled. New information can then be obtained from the resulting model, and this can then be used to modify the mechanism.
All mechanisms should start with a well-behaved starting model. This model should contain all the parts that will be used in the mechanism. In the case here, chymotrypsin, the essential parts are: (a) The protein chymotrypsin, (b) the substrate, and (c) the water molecule that will be consumed in the hydrolysis.
Building the first step, the starting model, requires a large amount of human effort. This has already been described for chymotrypsin, and will not be elaborated on here. It is sufficient to say that the starting model is valid.
In Step 2, the Ser195 hydroxyl oxygen forms a covalent bond with the peptide carbonyl carbon of Trp252; at the same time, the hydroxyl hydrogen migrates to His57. Simply moving the atoms towards each other and hoping that when the geometry is optimized they will be covalently bound is one option, but in most cases an operation of this type requires a good GUI and a lot of skill. Two strategies are recommended for modifying one stable system to form another stable system. First, the modification can be performed using JSmol. Three types of examples of the JSmol editor are: A simple hydrogen migration, an enol to keto reaction, and the formation of a tetrahedral intermediate. Operations of this kind are easy to perform using JSmol. The second method is to do everything using MOPAC. This is more difficult to set up, and takes considerably longer to run, but is much more versatile than JSmol.
A covalent bond can be made by using a reaction path to lower the interatomic distance between a pair of atoms that were initially well separated. This is done using a reaction coordinate, here the inter-atomic distance. Enzyme geometries are normally always expressed in Cartesian coordinates, but an interatomic distance requires internal coordinates. This means that one atom in the model must be defined in internal coordinates while all the other atoms are expressed in Cartesian coordinates. To set up the internal coordinate, look at the atom numbers of the two atoms that are to be moved together. The atom with the higher number will be the moving atom, atom A, the atom with the lower number will be the atom that the moving atom moves to, atom NA. Using a GUI, measure the distance between A and NA. Internal coordinates need an angle and a dihedral angle (a torsion angle), so select another atom to be used in defining an angle, atom NB and a dihedral, NC, and measure the angles A-NA-NB and A-NA-NB-NC. Edit the data set to replace the Cartesian coordinates of atom A with the new internal coordinates. Set the optimization flag to "-1" for the distance, and the optimization flags for the angle and dihedral to "1" An example of such an edited file would be:
N(ATOM 3483 N TRP C 252) 29.05500547 +1 2.39047640 +1 37.84548040 +1 C(ATOM 3484 CA TRP C 252) 27.71521548 +1 1.84347569 +1 38.05684432 +1 C(ATOM 3485 C TRP C 252) 2.508 -1 125.0 1 116.1 1 "OG SER G 195" "CB SER G 195" "CA SER G 195" ← Atom in internal coordinates O(ATOM 3486 O TRP C 252) 26.05974399 +1 0.53071662 +1 37.01434338 +1 (Columns lined up for ease of reading) C(ATOM 3487 CB TRP C 252) 26.71410326 +1 2.92780987 +1 38.49027948 +1
Here the carbonyl carbon in Trp 252, PDB label "C TRP C 252", is separated from the oxygen on Ser195, PDB label "OG SER G 195", by 2.508Å, and makes an angle of 125° with "CB SER G 195", and a dihedral of 116° with "CA SER G 195".
A useful alternative to using atom numbers when defining the connectivity is to use atom labels.
WARNING: A common problem occurs when the atom serial numbers in a PDB file are mistakenly used as atom numbers, and are then used for defining NA, NB, and NC. Serial numbers in a PDB file include non-atoms, for example a "TER" record (indicating the end of a chain) increments the serial number by one. Every time a "TER" is present the atom serial numbers for all subsequent atoms increases by one relative to the atom number. If you really want to use atom numbers, then to avoid this problem, run a 0SCF job using the system being studied. Use the output to identify the atom numbers to be used for NA, NB, and NC. (Tip: Check that the coordinate of the atom in the output file is the same as that in the GUI.)
Keywords for the reaction path can then be specified. In this case the keywords would be GNORM=30 STEP=-0.2 POINTS=5. This job is then run. This will take several hours. The normal reaction path accuracy is not needed, so the job can be run relatively quickly by using a large step and a large gradient criterion.
Edit the results to make a new data set. To replace the internal coordinates by Cartesian, change the distance optimization flag from "-1" to "1" and use keywords 0SCF XYZ. Then using a GUI, inspect the region of interest, here the region around atoms 3485 and 2737. In this specific case, the hydroxyl hydrogen is still attached to the Ser195 oxygen atom. Using the GUI, move this hydrogen atom so that it is ca. 1.0Å from the ring nitrogen of His57. The resulting structure is then run normally.
Sometimes it is necessary to break a covalent bond or to separate two atoms that are near to each other. If the resulting structure is not known, it is easier to simply increase the distance between the atom pairs. This is in contrast to the previous section, where two atoms are pulled together, and a bond may or may not break. In the case being described here, two atoms are pushed apart, and other bonds may or may not make.
An example of this type of process occurs in going from Step 3 to Step 4. In Step 3, there is an amine in close proximity to the tetrahedral carbon; the N - C distance being only 1.72 Å. In going from Step 3 to Step 4, the N - C distance is steadily increased. The "reaction path" would be defined on the keyword line by "GNORM=20 STEP=0.1 POINTS=15" A lower GNORM is used here, because the actual path is now of interest, not just the end product. The step-size is positive because the interatomic separation is increasing. 15 steps should be sufficient to increase the separation to ~ 3.2 Å.
In the data set, the nitrogen atom that is being moved is defined in internal coordinates, and the reaction coordinate is defined by the optimization flag "-1"
H(ATOM 3503 HZ2 TRP C 252) 25.32819248 +1 4.04224496 +1 43.95587272 +1
H(ATOM 3504 HZ3 TRP C 252) 24.19115878 +1 7.04202205 +1 41.07045984 +1
H(ATOM 3505 HH2 TRP C 252) 24.27393886 +1 6.22511542 +1 43.41277681 +1
N(ATOM 3506 N1 TRP C 252) 1.73 -1 101.26 1 81.74 1 3484 2737 2736 ← Atom in internal coordinates
C(ATOM 3507 C1 TRP C 252) 27.02440006 +1 -0.54493077 +1 34.80486955 +1
H(ATOM 3508 H1 TRP C 252) 28.28895525 +1 0.15309699 +1 36.31190994 +1
H(ATOM 3509 HG SER G 195) 28.37109739 +1 1.06634870 +1 34.92755644 +1
A better alternative to using atom numbers when defining the connectivity is to use atom labels.
If the only significant change is a proton shift or other very simple re-arrangement, then moving the atom "by hand" using a GUI is simplest. In going from Step 2 to Step 3, a proton, initially on His57, migrates to the nitrogen atom on Thr253. Using a GUI, identify the proton on His57, then move it so that it is near to the nitrogen of Thr253. It is helpful during this process to hide the rest of the model. Look at the His57 - Thr253 assembly from different orientations, to check that the proton is where it should be. Then, once everything looks correct, optimize the geometry of Step 3.