Provided that the three-dimensional structure of an enzyme is available, several
methods for predicting the selectivity and its stereochemical preference are possible
with various degrees of sophistication and effort [362–364]:
• The enzyme–substrate complex is constructed in its transition state for both
enantiomers and the energy value for both (diastereomeric) conformations
within the active site of the enzyme are calculated via molecular dynamics
(MD). The difference in free energy (ΔΔG
6 ¼ ) – obtained via force field calculations – yields semiquantitative results for the expected selectivity [365].
• The difference in steric interactions during a computer-generated approach of
two substrate enantiomers towards an acyl-enzyme intermediate can be used
instead [366].
• If the transition state is not known with some certainty, the substrate can be
electronically fitted into the active site of the enzyme (‘docking’). The orientation
of substrate enantiomers with respect to the chemical operator of the enzyme as
well as possible substrate movements can be analyzed via MD [367]. This is
achieved via (computer-generated) ‘heating’ of the substrate within the enzyme,
followed by a slow electronic ‘cooling process’, which allows the substrate
enantiomers to settle in their position representing the lowest energy minimum.
Because selectivities are determined by differences in free energy of transition
states, the first approach leads to the most accurate results. A simple free shareware program is AutoDock Vina [368]. However, it should be kept in mind, that
errors of up to Æ2.5 kcal/M in scoring energies are not uncommon, which renders
the estimation of stereoselectivities an educated guess at best (compare Table 1.4).
A representative example for the prediction of the stereochemical outcome of
the enzymatic hydrolysis of rac-1-phenylethyl acetate catalyzed by Candida antarctica lipase B based on its crystal structure (PDB: 5A71) is depicted in Fig. 2.10.
The bottom of the active site shows the catalytic triad consisting of Asp-187 and
His-224, which activate Ser-105 to perform a nucleophilic attack onto the carbonyl
group of the ester moiety, forming a tetrahedral oxy-anion intermediate (see also
Scheme 2.1). With the (R)-enantiomer (left), the bulky phenyl group is nicely
accommodated in the large lipophilic binding site (pink), while the small methyl
substituent is pointing upwards into the small pocket (yellow). In contrast, with the
(S)-enantiomer (right), the phenyl group would clash into the α-helix to the right,
while the methyl substituent would be inefficiently bound in the large pocket.
Hence, the prediction for the preferred enantiomer is (R), also denoted as
‘Kazlauskas-rule’ (Scheme 2.45).
2.1 Hydrolytic Reactions
81
methods for predicting the selectivity and its stereochemical preference are possible
with various degrees of sophistication and effort [362–364]:
• The enzyme–substrate complex is constructed in its transition state for both
enantiomers and the energy value for both (diastereomeric) conformations
within the active site of the enzyme are calculated via molecular dynamics
(MD). The difference in free energy (ΔΔG
6 ¼ ) – obtained via force field calculations – yields semiquantitative results for the expected selectivity [365].
• The difference in steric interactions during a computer-generated approach of
two substrate enantiomers towards an acyl-enzyme intermediate can be used
instead [366].
• If the transition state is not known with some certainty, the substrate can be
electronically fitted into the active site of the enzyme (‘docking’). The orientation
of substrate enantiomers with respect to the chemical operator of the enzyme as
well as possible substrate movements can be analyzed via MD [367]. This is
achieved via (computer-generated) ‘heating’ of the substrate within the enzyme,
followed by a slow electronic ‘cooling process’, which allows the substrate
enantiomers to settle in their position representing the lowest energy minimum.
Because selectivities are determined by differences in free energy of transition
states, the first approach leads to the most accurate results. A simple free shareware program is AutoDock Vina [368]. However, it should be kept in mind, that
errors of up to Æ2.5 kcal/M in scoring energies are not uncommon, which renders
the estimation of stereoselectivities an educated guess at best (compare Table 1.4).
A representative example for the prediction of the stereochemical outcome of
the enzymatic hydrolysis of rac-1-phenylethyl acetate catalyzed by Candida antarctica lipase B based on its crystal structure (PDB: 5A71) is depicted in Fig. 2.10.
The bottom of the active site shows the catalytic triad consisting of Asp-187 and
His-224, which activate Ser-105 to perform a nucleophilic attack onto the carbonyl
group of the ester moiety, forming a tetrahedral oxy-anion intermediate (see also
Scheme 2.1). With the (R)-enantiomer (left), the bulky phenyl group is nicely
accommodated in the large lipophilic binding site (pink), while the small methyl
substituent is pointing upwards into the small pocket (yellow). In contrast, with the
(S)-enantiomer (right), the phenyl group would clash into the α-helix to the right,
while the methyl substituent would be inefficiently bound in the large pocket.
Hence, the prediction for the preferred enantiomer is (R), also denoted as
‘Kazlauskas-rule’ (Scheme 2.45).
2.1 Hydrolytic Reactions
81