ABSTRACT
In silico computational chemistry constitutes an appealing method
to enhance our understanding of how enzymes work [1-5].
These tools provide researchers with a unique opportunity to
reveal details of catalysis at the molecular level that would be
difficult, or even impossible, to access by experimental methods
solely. Computational chemistry, in combination with laboratory
work, has become a cornerstone in today’s efforts to improve
promiscuous activities displayed by enzymes and to design novel
enzymes catalyzing hitherto unknown reactions [6-9]. The aim of
this chapter is to demonstrate the potential of using molecular
modeling to shed light on fundamental aspects of catalysis displayed
by amidases/proteases and esterases/lipases.a The potential of
using computer simulations to clarify the effects of introduced
mutations on enzyme catalysis was realized early on [10]. In silico
computational strategies for an elevated atomistic understanding of
enzymatic reaction mechanisms can be founded on [10-12]:
(i) Quantummechanics (QM)
(ii) Molecular dynamics (MD) and force-field methods
(iii) Hybrid methods (QM/MM [molecular mechanics])
(iv) Empirical valence bond (EVB) methods
Pioneering work on the basis of EVB calculations [13] and
MD simulations [14, 15] has paved the way for the development
of in silico methodologies into powerful and widespread tools
amenable for the study of enzyme catalysis. High-level first-principle
methods provide researchers with tools that can approach chemical
accuracy. However, even with the computational power available
today, QM calculations using a high level of theory are limited to
systems containing a couple of hundred atoms. Such small models
of enzymes have been termed “theozymes” [16]. Representing
enzymes by carving out carefully chosen active site models that are
used for quantum mechanical calculations has also been referred
to as the cluster approach [17]. In contrast to QM, MD simulations
are fast but chemical bonds are represented by springs. This
oversimplification of reality makes force-field-dependent methods
unable to represent bond-breaking and bond-forming processes.
In the hybrid QM/MM approach, the reacting atoms and a small
portion of the active site are treated quantum mechanically [18-
20] and capping atoms allow for interactions between the QM and
MD parts. The different theories listed above have their pros and
cons in achieving a trade-off between accuracy and speed in the
quest to increase our general understanding of how enzymes work.
Accessible time scales range from femtosecond for QM (i.e., bond
vibrations) tomillisecond forMD (i.e., conformational changes) [21].
