452
M. G. Quesne and S. P. de Visser
model) to 5.1 kcal mol
−1 . More importantly, the relative preference of the pathways
flips with the hydroxyl rebound barrier ( 4 TS R ) increasing from 5.1 kcal mol
−1 to
6.6 kcal mol
−1 , in the QM/MM model. This reversal of the barrier ordering leads
to a preference for olefin production that can be at least partly understood when
examining the 4 I R and 4 TS R structure shown in Fig. 6. As mentioned in previous
sections, the ability of a well-designed QM/MM protocol to replicate an accurately
solvated active site is one of the largest advantages of this type of technique, above
cluster models. The water network surrounding the iron(IV)-hydroxo of 4 I EA forms
a bridge to the guanidine groups of Arg 245 , which in turn was found to make the
rotation of hydroxyl group more difficult and thereby reversing the chemoselectivity
by pushing 4 TS R,EA above 4 TS D,EA . The authors subsequently highlight the fact that
P450 cam has a very tight-binding pocket that is likely to exclude water and therefore
prevent formation of analogous hydrogen-bonding networks and retain its preference
for forming the alcohol.
4.2 QM/MM Techniques Are Often Essential for Replicating
Minor Structural Anomalies that Lead to Large Changes
in Reactivity
Our second example in this section concerns AlkB, which is an α-ketoglutarate
(αKG)-dependent metalloenzyme that utilizes dioxygen and with the help of αketoglutarate forms a high-valent iron(IV)-oxo intermediate, which transfers its oxygen atom to the substrate [97–100]. AlkB is a member of another highly conserved
group of iron oxygenases that are responsible for the demethylation of methylated
DNA bases [101–103]. DNA bases are methylated in response to chemical damage
and if left unrepaired would eventually result in a halt in DNA replication by preventing Watson–Crick base pair formation [104–106]. AlkB is the proposed mechanism
based on experimental studies of taurine/α-ketoglutarate dioxygenase and computational modeling (mostly using QM cluster methods) is shown in Scheme 3. The
resting state is an iron(II) held to the protein via linkages with two histidine and
one carboxylate (Asp or Glu) ligands. Three water molecules complete the octahedral coordination environment of the resting state. It is envisaged that an initial
O 2 binding leads to the formation of an iron(III)-superoxo structure that attacks the
β-carbon of αKG, which is followed by decarboxylation to form succinate and CO 2
and an iron(IV)-oxo species [52, 107]. The latter reacts with substrate by hydrogen
atom abstraction and hydroxyl rebound steps, catalyzed by the resulting ferryl-oxo
intermediate species [101–103, 108].
However, while the crystal structure of the AlkB enzyme [109] confirms the
presence of a standard 2—His/1—Asp iron-binding motif, the oxygen coordination
site is surprisingly not proximal to the substrate (see Fig. 7), but orthogonal to it.
Therefore, a computational study was undertaken using a combination of QM/MM
and QM cluster techniques to assess the validity of the mechanism in Scheme 3
M. G. Quesne and S. P. de Visser
model) to 5.1 kcal mol
−1 . More importantly, the relative preference of the pathways
flips with the hydroxyl rebound barrier ( 4 TS R ) increasing from 5.1 kcal mol
−1 to
6.6 kcal mol
−1 , in the QM/MM model. This reversal of the barrier ordering leads
to a preference for olefin production that can be at least partly understood when
examining the 4 I R and 4 TS R structure shown in Fig. 6. As mentioned in previous
sections, the ability of a well-designed QM/MM protocol to replicate an accurately
solvated active site is one of the largest advantages of this type of technique, above
cluster models. The water network surrounding the iron(IV)-hydroxo of 4 I EA forms
a bridge to the guanidine groups of Arg 245 , which in turn was found to make the
rotation of hydroxyl group more difficult and thereby reversing the chemoselectivity
by pushing 4 TS R,EA above 4 TS D,EA . The authors subsequently highlight the fact that
P450 cam has a very tight-binding pocket that is likely to exclude water and therefore
prevent formation of analogous hydrogen-bonding networks and retain its preference
for forming the alcohol.
4.2 QM/MM Techniques Are Often Essential for Replicating
Minor Structural Anomalies that Lead to Large Changes
in Reactivity
Our second example in this section concerns AlkB, which is an α-ketoglutarate
(αKG)-dependent metalloenzyme that utilizes dioxygen and with the help of αketoglutarate forms a high-valent iron(IV)-oxo intermediate, which transfers its oxygen atom to the substrate [97–100]. AlkB is a member of another highly conserved
group of iron oxygenases that are responsible for the demethylation of methylated
DNA bases [101–103]. DNA bases are methylated in response to chemical damage
and if left unrepaired would eventually result in a halt in DNA replication by preventing Watson–Crick base pair formation [104–106]. AlkB is the proposed mechanism
based on experimental studies of taurine/α-ketoglutarate dioxygenase and computational modeling (mostly using QM cluster methods) is shown in Scheme 3. The
resting state is an iron(II) held to the protein via linkages with two histidine and
one carboxylate (Asp or Glu) ligands. Three water molecules complete the octahedral coordination environment of the resting state. It is envisaged that an initial
O 2 binding leads to the formation of an iron(III)-superoxo structure that attacks the
β-carbon of αKG, which is followed by decarboxylation to form succinate and CO 2
and an iron(IV)-oxo species [52, 107]. The latter reacts with substrate by hydrogen
atom abstraction and hydroxyl rebound steps, catalyzed by the resulting ferryl-oxo
intermediate species [101–103, 108].
However, while the crystal structure of the AlkB enzyme [109] confirms the
presence of a standard 2—His/1—Asp iron-binding motif, the oxygen coordination
site is surprisingly not proximal to the substrate (see Fig. 7), but orthogonal to it.
Therefore, a computational study was undertaken using a combination of QM/MM
and QM cluster techniques to assess the validity of the mechanism in Scheme 3