acyl transfer [650] and lactonization [651] reactions, which all proceed via a
tetrahedral carbanionic intermediate. In some instances, the reactions proved to
be enantioselective [652]. Elimination reactions [653, 654], reductions [655],
formation and breakage of C–C bonds [656, 657], ether cleavage [658], cis-trans
isomerization of alkenes [659] and even photochemical reactions [660] can be
catalyzed.
More importantly, pericyclic reactions, such as the Claisen rearrangement
[661, 662] (Scheme 3.56) or the Diels-Alder reaction [663, 664], have been
catalyzed by antibodies, which were raised against bicyclic and tricyclic
transition-state mimics. The latter reactions normally cannot be catalyzed by
enzymes. Also cationic cyclization reactions [665, 666] and an enantioselective
Robinson annulation were achieved [667].
The catalytic power of abzymes is most conveniently measured as the rate
acceleration compared to the uncatalyzed reaction. Recent studies have produced
artificial enzymes with catalytic activities approaching those of natural enzymes
[668], i.e. rate accelerations of abzymes are in the range of 10
3 –10
5 , and only in rare
cases they reach values of 10
8 . As yet, most enzymes are far better (~10
7 –10
17 ).
The data availble so far indicate that catalytic antibodies will certainly help us to
gain more insight into enzyme mechanisms but won’t playing an important role in
preparative biotransformations for the following reasons:
• They offer a unique access to artificial tailor-made enzyme-like catalysts and are
able to catalyze reactions which have no equivalent counterpart within the
diverse natural enzymes.
• The construction of abzymes having an opposite stereochemical preference is
possible.
• The catalytic efficiency is low in comparison to that of natural enzymes, because
abzymes were elicited against a model of the true transition state.
• The production of abzymes on scale is tedious [669] and ‘preparative-scale’
reactions are in the micromolar range [670, 671].
• The high binding energy for the substrate and product (up to 20 kcal/mol) causes
inhibition [672].
Computational De-Novo Design of Enzymes
Instead of a protein devoid of a chemical operator (an antibody), the active site
architecture of known proteins can be computationally re-designed to catalyze
non-natural reactions, by tapping on the ever growing number of high resolution
crystal structures [673].
14
In a first step, the transition state of a reaction is described using quantum
mechanical calculations. From this, an idealized active site of a protein bearing
functional groups positioned such as to maximize transition state stabilization is
computer-generated. This minimal active site description, called ‘theozyme’ or
14 Today, approximately 100,000 structure files are available from the Protein Data Bank and about
a dozen are added every day.
388
3 Special Techniques
tetrahedral carbanionic intermediate. In some instances, the reactions proved to
be enantioselective [652]. Elimination reactions [653, 654], reductions [655],
formation and breakage of C–C bonds [656, 657], ether cleavage [658], cis-trans
isomerization of alkenes [659] and even photochemical reactions [660] can be
catalyzed.
More importantly, pericyclic reactions, such as the Claisen rearrangement
[661, 662] (Scheme 3.56) or the Diels-Alder reaction [663, 664], have been
catalyzed by antibodies, which were raised against bicyclic and tricyclic
transition-state mimics. The latter reactions normally cannot be catalyzed by
enzymes. Also cationic cyclization reactions [665, 666] and an enantioselective
Robinson annulation were achieved [667].
The catalytic power of abzymes is most conveniently measured as the rate
acceleration compared to the uncatalyzed reaction. Recent studies have produced
artificial enzymes with catalytic activities approaching those of natural enzymes
[668], i.e. rate accelerations of abzymes are in the range of 10
3 –10
5 , and only in rare
cases they reach values of 10
8 . As yet, most enzymes are far better (~10
7 –10
17 ).
The data availble so far indicate that catalytic antibodies will certainly help us to
gain more insight into enzyme mechanisms but won’t playing an important role in
preparative biotransformations for the following reasons:
• They offer a unique access to artificial tailor-made enzyme-like catalysts and are
able to catalyze reactions which have no equivalent counterpart within the
diverse natural enzymes.
• The construction of abzymes having an opposite stereochemical preference is
possible.
• The catalytic efficiency is low in comparison to that of natural enzymes, because
abzymes were elicited against a model of the true transition state.
• The production of abzymes on scale is tedious [669] and ‘preparative-scale’
reactions are in the micromolar range [670, 671].
• The high binding energy for the substrate and product (up to 20 kcal/mol) causes
inhibition [672].
Computational De-Novo Design of Enzymes
Instead of a protein devoid of a chemical operator (an antibody), the active site
architecture of known proteins can be computationally re-designed to catalyze
non-natural reactions, by tapping on the ever growing number of high resolution
crystal structures [673].
14
In a first step, the transition state of a reaction is described using quantum
mechanical calculations. From this, an idealized active site of a protein bearing
functional groups positioned such as to maximize transition state stabilization is
computer-generated. This minimal active site description, called ‘theozyme’ or
14 Today, approximately 100,000 structure files are available from the Protein Data Bank and about
a dozen are added every day.
388
3 Special Techniques