Now that the results obtained for all 16 truncated models of AChE’s active site have been presented, they can be discussed. The first element of note is that the inclusion of the acetate in the model systematically prevented the reaction. It was only possible to optimize a product structure for the minimal model but no transition state could be optimized. In those models, which mimics Glu334, the reactivation of VX-inhibited AChE with 2-PAM appears to
be very endothermic with energy barriers at least superior to 30 kcal.mol-1. This is surprising
as imidazole has a pKa of 7.0, higher than the pKa of 4.8 of the acetate. This difference in pKa
should mean that the imidazolium should keep the proton over the acetate. However, classical
pKa values are measured in bulk solvent, and considering that the imidazolium forms a
hydrogen bond with the VX-Ser203 adduct model, the acidity of the imidazolium might be increased over that of acetic acid. When the implicit solvent model COSMO was used for the minimal model with the acetate mimicking Glu334, the reaction was able to take place. A product was optimized despite the large endothermicity. It could explain the preventing of the reaction by the acetate in the gas phase as being due to a lack of compensation of the charge of the acetate. In the forthcoming QM/MM simulations, the acetate will be naturally compensated by a well described enzymatic environment.
For all the models that did not include an acetate the reaction could be simulated. In all eight cases where the acetate was not included, the reaction has a single transition state and no stable pentavalent intermediate (see Scheme VI-1 and Figure I-3). Thus, in all those
steps the mechanism is an SN2. This was unexpected as all other investigations of AChE’s
reactivation using a similar truncated model methodology indicated an addition-elimination
mechanism.[7] When the reactivation was modelled in the minimal model in water solvent
implicit solvation using COSMO, the reactivation still has a SN2 mechanism but with an
addition elimination character. The other truncated model investigations of AChE’s
reactivation were made with tabun-inhibited AChE instead of VX-inhibited AChE. [7] The steric
hindrance caused by the N-dimethyl substituent of tabun might explain the addition- elimination mechanism.
Scheme VI-1. Pentavalent intermediate for the addition-elimination mechanism of VX-inhibited AChE
reactivation by an oxime.
[7] a) R. Lo, B. Ganguly, Mol. BioSyst. 2014, 10, 2368–2383. b) R. Lo, N. B. Chandar, M. K. Kesharwani, A. Jain, B. Ganguly, PLoS One 2013, 8, e79591.c) J. Wang, J. Gu, J. Leszczynski, M. Feliks, W. A. Sokalski, J.
In Figure VI-1 the values for both key distances P-Ox and P-Oser203 of the reactants, products, and transition states are presented for the different truncated models. What is apparent is that the transition state is systematically closer in structure to the reactant than to the product. This early transition state is generally (to the exception of benzene) even earlier for models including the oxyanionic hole. The formation of the transition state from the reactant involves very little change in the P-OSer203 distance forming an almost straight line in Figure VI-1. In that respect, it evokes what could be the addition transition state for a hypothetic addition elimination mechanism.
Figure VI-1. Reactive distances for all the optimized reactant product and transition state structures for all models
that did not include acetate. Models including the oxyanionic hole are indicated by diamonds and models without the oxyanionic hole are indicated by circles. For reactants, the shapes are filled by bands, for the transition states the filling is plain, and for the products, the filling is dotted.
If the mechanism truly is an addition-elimination, a pentavalent intermediate should be found. It is not the case according to the calculations presented in this chapter. It could mean that the stabilization brought by the oxyanionic hole to the product state is greater than the stabilization it brings to the reactant state. It can be viewed as a case of a dissymmetric three state VB configuration mixing diagram where the products state rests lower in energy
compared to the pentavalent state (Figure VI-2).8
Figure VI-2. VB configuration mixing diagram from the reactivation.
The simulations confirm the role of the oxyanionic hole in the process of reactivation in the stabilisation of transition states.[9] As can be seen in Table VI-1, the inclusion of the
oxyanionic hole systematically reduces the energy of the transition state compared to similar models without the oxyanionic hole. The hydrogen bonds the oxyanionic hole forms with the oxygen involved in the phosphoryl bond reduce the energetic cost of over-polarizing the phosphoryl bond. The energy difference of the reactivation is also decreased by the oxyanionic hole in all four cases. The oxyanionic hole appears to stabilize the product more than it does the reactant. In the models including the oxyanionic hole, the reactivation is always exothermic (see Figure VI-3).
Table VI-1. Changes in reaction barrier (ΔE*) and energy difference (ΔΔE) as a result of the inclusion of the
oxyanionic hole in various models. All energies in kcal.mol-1. ΔE* (kcal.mol-1)
ΔΔE (kcal.mol-1)
Minimal model -2.93 -4.66
Model with benzene -1.26 -4.32
Model with phenol -0.73 -4.89
Model with indole -1.18 -1.65
[9] a) N. Qian, I. M. Kovach, FEBS Lett. 1993, 336, 263–266. b) Y. Li, L. Du, Y. Hu, X. Sun, J. Hu, Can. J. Chem.
Figure VI-3. Reaction profiles for the reactivation in various truncated models. Energies were obtained with
B3LYP/def2-SV(P) and are in kcal.mol-1.
Figure VI-3 can be used as a basis for the discussion of the influence of the aromatic
rings in the reactivation. Their inclusion in the model, with or without the oxyanionic hole systematically stabilizes the product more than the reactant. The increase in exothermic character of the reactivation when the oxyanionic hole is not identical for all aromatic rings.
The exothermicity is increased by 4.89 kcal.mol-1 for phenol but only by 1.65 kcal.mol-1 for
indole. This difference might be due to interactions between the aromatic rings and the oxyanionic hole that modify the way the oxyanionic hole interacts with the 2-PAM-VX and Ser203-VX adducts. The effect of aromatic rings on the energy barrier is inconsistent as benzene and indole increase the energy barrier while phenol decreases it. It could be concluded that phenol has a stabilizing interaction with 2-PAM but all activation energies for
aromatic rings are within 3 kcal.mol-1 which falls in the limits of the truncated model