Molecules 2020, 25, 2064
17 of 22
were not expected to take place. However, the computations could not predict which of the exothermic
reactions did actually take place since this is also dependent on the environment within the enzyme and
the geometry of the non-covalent enzyme-inhibitor complex. Computations that take these influences
into account, e.g., QM/MM or even QM/MM/MD computations, are only possible when the necessary
crystallographic information of the enzyme–inhibitor complex are available, which was not the case
for the systems presented herein.
The reaction energies obtained from the measured k and k values for the nitrile-compounds
2
3
4
and
4
were between
−
0.89 and 1.5 kcal/mol. These values fit best to the value for the attack at the
−
nitrile group computed with SCS-MP2 but the formation of III a or IV a were predicted to be more
exothermic, i.e., they should be formed. However, within an enzyme environment they might become
less favored due to two reasons. Firstly, the orientation of the inhibitor within the active site might
only allow an attack at the nitrile. Additionally, the formation of III or IV might be hampered by
steric effects because the flat sp hybridized inhibitor is transformed into a considerably more bulky
form due to the formation of two sp centers. Nevertheless, the computations definitely excluded the
2
3
formation of intermediates, which should lead to an irreversible enzyme inhibition. This was in line
with the experimental results. It is also important to note that the formation of IV b, which was not
observed in the solvent experiment, was definitively excluded by our computations.
The differences between SCS-MP2 and DFT approaches were even more obvious for the
chloro-substituted 1,4-naphthoquinone (Table 5). As for the nitrile compound, SCS-MP2 predicts that the
formation of compounds III a 1 and its enantiomer III a 4 were most favorable (∆Ereac = −17.3 kcal/mol).
Additionally, as computed for the nitrile compound, other compounds could be formed because their
reactions energies differed only slightly from those of the formation of III a 1 and III a 4. However,
the products that indicate an irreversible substitution reaction (III b and IV b), were again too high
in energy to be formed. In contrast, most DFT functionals predicted the formation of compound
+
−
IV b, i.e., the addition to the double bond followed by the elimination of HCl (H + Cl ). Since the
elimination product (i.e., chloride) is expected to diffuse away and because it is not sufficiently reactive
for a reverse reaction, most functionals predicted an irreversible reaction. This was in agreement with
the experimental findings for the reaction of the chloro-derivative with excess LMW thiol in solution in
the presence of a base. The product of this solution reaction corresponds to product IV b in which the
chloride was substituted by the attacking thiol group. Does this result indicate that the DFT functionals
are right while SCS-MP2 is wrong? This is not the case. For III a 1, SCS-MP2 predicted a reversible
reaction, i.e., reactants and products were in equilibrium. The formation of IV b was also predicted
to be exothermic, i.e., this reaction would also take place. Due to the difference in the exothermicity,
IV b was formed to a lower extent, but because this reaction is irreversible while the formation of III a 1
is reversible, finally only the product of the irreversible reaction would be found. For the enzyme
environment, both reactions can be expected to be in equilibrium because the chloride as elimination
product will be formed within the active site of the enzyme, i.e., its diffusion will be strongly hindered.
In this case, the formation of e.g., III a 1 would be considerably favored so that the irreversible reaction
will be suppressed. It is also important to note that our computations are in line with the experimental
observation that in the solvent reaction of an LMW thiol with the chloro-substituted compound the
elimination product IV b was formed while it was not formed for the CN-substituted one.
For the chloro-substituted compound, the experimental and computed reaction energies differed
considerably. This may be owed to the fact that the computations neglected steric as well as electronic
effects arose from the enzyme environment.
According to the QM computations, for both compound classes, namely the chloro and the
nitrile derivatives, the reversible attack of the thiolate at the
α
- or
β-position of the double bond
(yielding products III a or IV a) should be possible. The reaction with a LMW thiol took place at
the
α
-C atom in case of the chloro derivative (yielding IV b). The experimentally determined
) were in the same range for both enzymes
1.5 kcal/mol, see Table 2). Taken together, it may be hypothesized that for at least the
∆E
reac
values for both the chloro and the nitrile derivatives (
0.85 to
2–4
(−
−