Chapter 3 Quantum cluster models of phosphorylation and
3.6 The impact of restraints on quantum cluster
3.6.2 Dephosphorylation system
Parathion and dichlorvos have dimethyl side chains, therefore the
phosphoadduct has dimethyl side chains, whereas dECP and diazinon
have diethyl side chains, and therefore the phosphoadduct has diethyl
side chains. For this reason, the dephosphorylation study would be made
of two systems instead of four. In order to have more data points for
comparison, two new dephosphorylation systems, the E3-VX(R) and E3-
VX(S) adducts, were introduced. For these, VX was docked to E3 (see
Chapter 4) and manually moved to a distance suitable for attack.
During examination of the converged structures of the dephosphorylation
system for the diethyl and dimethyl substrates it was noticed that in TS2
the His471 residue moves closer to the Ser218, in order to transfer a proton
to the serine oxygen atom. This contact still exists, as a hydrogen bond, in
the intermediate. There is no such hydrogen bond, however, in the X-ray
structure of the free enzyme. It is also not necessary in TS1, and it is
therefore also absent from the intermediate as determined by following
the IRC from TS1. Hence, two different intermediate geometries with
different energies are obtained from IRC of TS1 and TS2.
In the case of the diethyl substrate, the barriers initially obtained were 52
kJ/mol for the first step of the reaction and 60 kJ/mol for the second. It is
worth noting that, because of the occurrence of different intermediate
geometries as described in the previous paragraph, calculating ΔE2 using
the intermediate obtained from TS1 resulted in an erroneous negative
barrier, which means this intermediate geometry has higher energy than
the transition state. Because of this error, the position of His471 in the
initial structure used for TS1 search was changed to match its position in
leads to a geometry of the intermediate that is virtually identical to that
obtained by IRC from TS2 and which therefore has its same energy.
Furthermore, reoptimisation of the reactant was attempted by IRC from
TS1
1and the geometry obtained substantially decreased the calculated
electronic energy barrier, from 52 kJ/mol (this number is shown in
brackets in Table 3.9) to 22 kJ/mol (see Figures 3.15 and 3.16). In this case,
the optimisations based on modified geometries allowed for a whole
reaction path to be constructed. This highlights the need for special care in
quantum cluster calculations, in which there exists a complex PES with
multiple possible pathways.
For the dimethyl substrate, the first set of barriers obtained were ΔE1 = 26
kJ/mol and ΔE2 = 14 kJ/mol. The intermediates obtained by IRC from TS1
and TS2 were different, and calculating
ΔE2 using the intermediate
obtained from TS1 yielded and erroneous value of -41 kJ/mol. As
discussed for the diethyl adduct before, this error occurs because the two
halves of the reaction path do not match. Therefore, an optimisation of TS1
was carried out with a modified geometry in which His471 was moved to
its position in TS2. The subsequent IRC calculation and optimisation of the
intermediate, however, led to a higher energy barrier for the second step
(33 kJ/mol, shown in brackets in table 3.9) than that obtained when the
intermediate is optimised from TS2 (14 kJ/mol). The barrier for the first
step of the reaction also increased, from 26 kJ/mol to 44 kJ/mol when this
new TS1 geometry was used to obtain a new reactant geometry. Therefore,
in the case of the dimethyl adduct, the geometries used to try to improve
the path led to worse results instead. The time available for this project did
not make it possible to attempt more path searches for this system, and a
continuous path is yet to be found. Yet another problem faced with the use
of restraints was that in one case, namely the dephosphorylation of the E3-
VX(R) adduct, they led to artificial ‘negative’ barriers, which means that
the geometry of the enzyme needs to relax in order to produce realistic
energies for TS and minima. For a comparison of the geometries produced
by different substrates, see Figures 3.17 – 3.20.
Figure 3.15.Effect of the position of His471 on the geometry of TS1 and reactant
on the dephosphorylation geometries of dimethyl adducts.
A. Reactant obtained
via IRC from the TS1 geometry initially optimised (magenta) and the
reactant obtained from the modified TS1 geometry (green). B. Geometry of
TS1 optimised initially (magenta) and TS1 reoptimised with a modified
geometry (see text for details) (green).
Figure 3.16.
Effect of the position of His471 on the geometry the intermediate
geometry on the dephosphorylation geometries of dimethyl adducts.A. Geometry
of the intermediate obtained via IRC from TS2 (magenta) and geometry
obtained from TS1 initially optimised (blue).
B. Geometry of the
intermediate obtained via IRC from TS2 (magenta) and geometry obtained
from TS1 reoptimised with a modified geometry (see text for details)
(green).
R1
R2
ΔE1
ΔE2
ΔE1
ΔE2
Diethyl adduct
22[52]
60
n/a
n/a
Dimethyl adduct
26[44]
14[33]
7
57
VX(R) adduct
n/d
-27
-10
n/d
VX(S) adduct
24
n/d
n/d
24
Table 3.9.
Dephosphorylation reactions with different substrates and restraints.
Electronic energy (RI-MP2/aug-cc-pvtz) barriers for the four
dephosphorylation reactions studied. Restraints were R1 (freezing the first
carbon after the functional group of a residue) or R2 (restraining all but
the substrate to the geometry already established for the reaction that has
dECP as a substrate). R2 was not applicable (n/a) to the dECP system. For
several systems transition states were not determined (n/d) as they were
not located after several attempts. All values are in kJ/mol. Numbers in
brackets indicate a less-favourable energy barrier obtained with other
geometries (as described in the text).
Figure 3.17.
Dephosphorylation reaction, reactant geometries.
A. E3-diethyl
adduct (blue) and E3-dimethyl adduct (green). B. E3-diethyl (blue) and
E3-VX(S) adduct (purple). C. Overlap of all the systems (dimethyl, diethyl
and E3-VX(S)). The geometries used in this and all other captions, unless
otherwise specified, are those that lead to a lower energy barrier, as these
would be the preferred conformations.
Figure 3.18.
Dephosphorylation reaction, TS1 geometries.
A. Diethyl (blue)
and dimethyl (green). B. Diethyl (blue) and VX(S) (purple). C. Overlap of
Figure 3.19. Dephosphorylation reaction, intermediate geometries. A. E3-diethyl
adduct (blue) and E3-dimethyl adduct (green). B. E3-diethyl (blue) and
E3-VX(R) adduct (purple). C. Overlap of these intermediate structures.
Figure 3.20. Dephosphorylation reaction, TS2 geometries. A. E3-diethyl adduct
(blue) and E3-dimethyl adduct (green). B. E3-diethyl (blue) and E3-VX(R)
adduct (purple). C. Overlap these TS2 geometries.
In document
Computational studies of the E3 carboxylesterase from Lucilia cuprina
(Page 89-93)