172 COMPUTER HARDWARE, SOFTWARE
originate from the exact solution of an idealized many - electron problem,
mainly an electron gas of constant total electron density. In practice, one
establishes functional relationships between the exchange and correlation
energies for the “ idealized gas ” and the total electron density. However, such
relationships are not unique and do not as yet lead to systematic progression
to a low - energy structure. Practical density functional calculations make use
of explicit atomic basis sets as described for ab initio methods. Additionally,
numerical integration steps are necessary. These may lead to loss of precision
and slower calculations compared to ab initio methods, especially for small
molecules. Costwise density function methods are particularly useful for large
systems when compared to other Hartree – Fock methods. In Section 4.5,
density functional methods are described for high - end programs, usually run
on workstation - level computers.
Application of density function methods to bioinorganic systems was the
subject in the American Chemical Society Inorganic Section at the August
2000 national meeting. In introductory remarks, David A. Dixon of the Pacifi c
Northwest National Laboratory stated that density functional theory has been
shown to be an effective way — in accuracy and computational cost — to predict
structures and vibrational spectra of inorganic compounds in comparison to
other methods. He stated that DFT can be used to predict other properties
including reaction energies, electron densities, excited - state spectra, and NMR
chemical shifts. Subsequent speakers in the section told of applying localized
perturbation approaches and density functional theory to systems with hun-
dreds of atoms, allowing accurate calculations including electron correlation
to be carried out. One researcher, M. B. Hall, reported density functional cal-
culations on models for [Fe] - hydrogenase. He related these to structures and
vibrational frequencies of the observed redox forms of the enzyme, as well as
to the reaction mechanism at the enzyme ’ s diiron active center.^30 R. A. Friesner
spoke on large - scale ab initio quantum chemical calculations on biological
systems.^31 More recently, Friesner and Guallar reported a hybrid quantum
mechanics/molecular mechanics study of the cytochrome P450 CAM enzymatic
catalysis cycle.^32 Cytochrome P450 CAM , a monooxygenase, is discussed in
Section 7.4 (see especially Section 7.4.2 and Figures 7.12 and 7.13 ), and the
catalytic cycle is shown in Figure 7.14. The cytochrome P450 CAM enzyme, one
of the most studied P450s, catalyzes the oxidation of camphor (see Figure
7.12). The reference 32 study ’ s results showed an active role for cytochrome
P450CAM in several catalytic steps. The enzyme is involved in controlling the
energy gap between high - and low - spin states in the substrate binding process.
At the enzyme ’ s active site, selective interaction of thr252 and the distal dioxy-
gen molecule causes O – O bond cleavage. The protein environment was found
to catalyze substrate (camphor) hydrogen atom abstraction with a very low
free - energy barrier ( ∼ 8 kcal/mol). The reference 32 authors believe that their
results provide an explanation as to why the proposed catalytically competent
intermediate — a ferryl iron – oxo species, Fe IV =O, also called “ compound I ” —
cannot be trapped experimentally.