next up previous contents
Next: Semiempirical calculations Up: Mandelate Racemase enzyme Previous: Tables of results   Contents


Quantum calculations on mandelate mimic systems

An easy question and a difficult answer:
This thesis will have a significant part devoted to the location of stationary points in big enzymatic systems. However, looking to the current publications in theoretical chemistry and mainly quantum chemistry, we find many studies where the enzyme is simplified to a few tens of atoms [206]. Then, an obvious question emerges. Will we obtain the same results if we model the enzyme by reducing the system to a small representation of the active site?

This strategy when applied to enzymes is widely used, the model is usually referred as biomimetic, cluster or simply gas phase model. However the reduction of the studied system from several thousands to a few tens of atoms has some important drawbacks that must be accepted from the beginning:

Energy:
The absence of the whole environment excludes the important long range interactions. Some authors solve this problem with a continuum model.
Mechanism:
The oversimplification of the active site may exclude some a priori unknown residues that have an important role in the mechanism 3.11.
Artificial constraints:
In order to keep an adequate structure of the active site these calculations usually need to constrain some selected degrees of freedom. Commonly, some atoms are frozen at the position found in the PDB structure. However, the selection of atoms to freeze and their fixed coordinates may change along the reaction steps of a mechanism.
Unreal structures:
Despite of the constriction of some atoms we may obtain unreal interactions between two residues due to the oversimplified structure.
Obviously, the two last points will depend on how flexible is our system. In any case, gas phase modelizations are still useful in some chemical systems where accurate potential energies and a strict control of electronic state are essential (mainly in metalloproteins) [206,207].

In this section we model the racemization reaction of mandelate substrate already studied in section 2.3.3. Two gas phase models have been chosen. The first one (model 1) is the quantum part selected in the previous QM/MM study. In figure 2.12 we show again the 88 atoms represented in model 1.

Figure 2.12: Gas phase model 1 of the active site of Mandelate Racemase for Quantum Mechanics calculations.
\includegraphics[width=\textwidth]{Figures/Gas/model1.eps}

In the second model (model 2; 60 atoms) we excluded the magnesium atom and the corresponding ligands. As it is shown in figure 2.13, we only include the mandelate substrate, the two general acid-base residues Lys166 and His297 and the catalytic residues Lys164 and Glu317.

Figure 2.13: Gas phase model 2 of the active site of Mandelate Racemase for Quantum Mechanics calculations.
\includegraphics[width=\textwidth]{Figures/Gas/model2.eps}

We have performed calculations at the PM3-SRP(Mg) semiempirical level and at the B3LYP/6-31G* density functional level of theory with the two different models. The calculations have been carried out by a modified version of Gaussian 98 package[239] to incorporate the SRP parameters for Mg atom in PM3 (model 1).


Subsections
next up previous contents
Next: Semiempirical calculations Up: Mandelate Racemase enzyme Previous: Tables of results   Contents
Xavier Prat Resina 2004-09-09