344
Z. Fu and J. Chen
17.2 Method and Materials
With the fast advancement of quantum chemical theories and high-performance
computing capacities, quantum chemical methods such as density functional theory (DFT) calculations can now accurately reproduce the electronic configuration
and properties of various systems and have demonstrated advantages in predicting
chemical reactivity. However, DFT methods are restricted by the system size, and
converged results obtained on a system with >200 atoms are less reliable [11]. Therefore, common practice is simulating P450 enzymes by adopting a simplified model
of the active center (Cpd I in Fig. 17.4), also known as a cluster model in DFT calculations. The simplification is done by truncating the adjacent substituents (Fig. 17.2)
to protoporphyrin and replacing the vertical cysteine residue with –SH, –SCH 3 , or
–SCys. The enzyme environment is then mimicked with implicit solvation models, e.g., the polarizable continuum model (PCM). The basic idea for probing P450
catalyzed reactions using the cluster model lies in a general fact: Although P450
enzyme structures vary with species and isoforms, they share analogous active centers that are intrinsically responsible for their metabolic reactions. Thus, simulating
the active center with relatively accurate quantum chemical methods would serve
to effectively answer chemistry-related questions such as reaction conformations,
electronic structures, and reactivity.
As the computing capacity improves, cluster models can better describe more
complex systems by incorporating sufficiently large numbers of atoms. Nevertheless,
small models based on the active site retain their superiority in dealing with reaction
mechanisms at the early stage. Firstly, small models are suitable for quick probing of
various reaction routes because of a low computational cost. Secondly, employment
of simple models avoids artifacts and tends to receive more accurate results. One
general rule in computational biological chemistry is that when a large discrepancy
occurs between large and small model results, results from the smaller models are
more likely to be correct [11]. It remains difficult to obtain computationally correct
results for simulations using large models.
Harris et al. [12] first investigated the electronic structures for the resting state of
P450cam, an isoform which specifically binds with camphor, using combined quantum chemical Hartree–Fock and molecular dynamics (MD) calculations. Consistent
with the electron spin echo envelope modulation (ESEEM) spectroscopic data, a
low-spin doublet state was characterized for the resting state, which is stabilized
by the electrostatic interactions with residues surrounding the active site and the
ligated water molecule. Subsequent work by Shaik et al. [8] further probed the electronic structures and properties of other intermediates in the P450 catalytic cycle and
unveiled several key factors that determine the catalytic reactivity of P450 enzymes.
One factor concerns the donor ability of the –SH substituent, also termed “push
effects,” since the Fe–S distance in Compound I would affect the electronic configuration and thus its oxidative capability. Secondly, the protonation mechanisms
of Compound 0 would decide the productivity of Compound I, in a way that an
ineffective protonation at the proximal oxygen atom would possibly activate O 2 into
Z. Fu and J. Chen
17.2 Method and Materials
With the fast advancement of quantum chemical theories and high-performance
computing capacities, quantum chemical methods such as density functional theory (DFT) calculations can now accurately reproduce the electronic configuration
and properties of various systems and have demonstrated advantages in predicting
chemical reactivity. However, DFT methods are restricted by the system size, and
converged results obtained on a system with >200 atoms are less reliable [11]. Therefore, common practice is simulating P450 enzymes by adopting a simplified model
of the active center (Cpd I in Fig. 17.4), also known as a cluster model in DFT calculations. The simplification is done by truncating the adjacent substituents (Fig. 17.2)
to protoporphyrin and replacing the vertical cysteine residue with –SH, –SCH 3 , or
–SCys. The enzyme environment is then mimicked with implicit solvation models, e.g., the polarizable continuum model (PCM). The basic idea for probing P450
catalyzed reactions using the cluster model lies in a general fact: Although P450
enzyme structures vary with species and isoforms, they share analogous active centers that are intrinsically responsible for their metabolic reactions. Thus, simulating
the active center with relatively accurate quantum chemical methods would serve
to effectively answer chemistry-related questions such as reaction conformations,
electronic structures, and reactivity.
As the computing capacity improves, cluster models can better describe more
complex systems by incorporating sufficiently large numbers of atoms. Nevertheless,
small models based on the active site retain their superiority in dealing with reaction
mechanisms at the early stage. Firstly, small models are suitable for quick probing of
various reaction routes because of a low computational cost. Secondly, employment
of simple models avoids artifacts and tends to receive more accurate results. One
general rule in computational biological chemistry is that when a large discrepancy
occurs between large and small model results, results from the smaller models are
more likely to be correct [11]. It remains difficult to obtain computationally correct
results for simulations using large models.
Harris et al. [12] first investigated the electronic structures for the resting state of
P450cam, an isoform which specifically binds with camphor, using combined quantum chemical Hartree–Fock and molecular dynamics (MD) calculations. Consistent
with the electron spin echo envelope modulation (ESEEM) spectroscopic data, a
low-spin doublet state was characterized for the resting state, which is stabilized
by the electrostatic interactions with residues surrounding the active site and the
ligated water molecule. Subsequent work by Shaik et al. [8] further probed the electronic structures and properties of other intermediates in the P450 catalytic cycle and
unveiled several key factors that determine the catalytic reactivity of P450 enzymes.
One factor concerns the donor ability of the –SH substituent, also termed “push
effects,” since the Fe–S distance in Compound I would affect the electronic configuration and thus its oxidative capability. Secondly, the protonation mechanisms
of Compound 0 would decide the productivity of Compound I, in a way that an
ineffective protonation at the proximal oxygen atom would possibly activate O 2 into