Use of the LOCATE-TS technique to locate a transition state

LOCATE-TS is ideally suited for finding transition states in enzyme mechanisms.  Instead of several worked examples involving different enzymes, the examples here all refer to different steps in the chymotrypsin catalyzed hydrolysis of a peptide bond.  For a detailed description see the published article. To reproduce the chymotrypsin mechanism, see the supplementary materials

General notes

A good starting model is essential for investigating enzyme mechanisms. This model should represent the first step in the catalytic cycle.  Check it carefully to ensure that it is as perfect as possible - if a change has to be made later on, it might require all previous steps to be changed and re-run.  Although this is not very time-consuming, it is annoying, and can usually be avoided by having a better starting model.

Make intermediates by moving atoms around using a GUI.  If necessary, the sequence of the atoms can be changed, but do not change the part of the PDB atom label that identifies the atom.  For example, if the MOPAC data-set line for an atom consists of: "C(ATOM 36 CA PRO E 4) 8.35040525 +1 13.76512201 +1 27.02235000 +1" the part that should not be changed is "CA PRO E 4".  Within a catalytic cycle, an atom might change from being covalently bonded to one specific atom to being covalently bonded to another atom, that is, it is involved in a chemical reaction.  Although the atom is now in a different environment, its PDB label must not be changed. This means that, at the end of the cycle, some atoms will appear to have the wrong labels.  This is expected and is necessary. An incidental benefit of this requirement is that the path of an atom through the catalytic cycle can easily be followed.  If there is a need to update the atom labels, use keyword RESIDUES, but if tat is done, the resulting structure should not be used in any runs involving keywords LOCATE-TS, SADDLE or COMPARE.

From the "distance" between adjacent intermediates it might appear that they have very different geometries.  This is a common phenomenon, caused by the large motion of the enzyme as it settles into the energy minimum for the intermediate.  That this large "distance" is not important can easily be demonstrated by using LOCATE-TS=(C:5,5;), this moves both systems to near the bottom of the reaction barrier, to the point where the slope amounts to 5 kcal.mol^-1 .Å^-1. At that point, the "distance" should be much smaller than it was.

There is an increased chance of user-error in preparing LOCATE-TS jobs, so the addition of keywords that trap errors is worthwhile.  Always add keyword CHARGE=n, even when the system has a net charge of zero.  Of course, check that the value of "n" is correct before starting a LOCATE-TS job! Do not use GEO-OK or LET unless necessary - these keywords are intended to trap errors in data.