456
M. G. Quesne and S. P. de Visser
strate approach for the correct orbital overlap is approximately 180°, which usually
results in less stereochemical clashes than seen in the optimal 120° of the π-pathway
and also leads to the maximum level of exchange interactions between unpaired electrons on Fe(III) [52, 113, 116–121]. The σ -pathway for hydrogen-transfer barrier
(
5 TS H, σ ) included a Fe–O–C angle of 141.4° which is obviously considerable smaller
than the optimal; while the angle of substrate approach seen in
5 TS Hπ (131.4°) is
considerably more favorable for π hydrogen abstraction. The potential energy surfaces produced by both pathways (from 5 R to 5 I) are shown in Fig. 8 and illustrate
that both
5 TS H,σ and
5 TS H,π have comparable energies: 18.2 versus 18.6 kcal mol
−1 ,
which is considerably more realistic then the value of over 20 kcal mol
−1 found using
cluster models.
The final step of the reaction mechanism involves a rebound of the hydroxyl group
back onto the radical substrate intermediate, with both
5 I σ and
5 I π forming strong
hydrogen-bonding interaction between carboxylate group of Asp 133 and the Fe(III)
− OH. This then leads to relatively high rebound barriers of >13.1 kcal mol
−1 ,
which are very far away from the 0.6 kcal mol
−1 found with the DFT cluster model
calculations [115].
5 Conclusion
Computational studies on non-heme and heme monoxygenases and dioxygenases
are useful for the understanding of metalloenzymes and in particularly for providing
insights into the chemical and physical properties of short-lived and highly reactive
intermediates. The key to these studies remains the choice of the model, which should
be a good representation of the real system. Two approaches have been used, namely
QM models and QM/MM techniques. The latter take the structure of the full enzyme
with a solvent layer into consideration whereas the former only include key features
of the first and second coordination sphere. We have shown several examples where
small active site models using the QM cluster technique give viable results that
compare well with experiment. For systems, however, where substrate binding is
tight and or a network of hydrogen-bonding interactions exist, a complete QM/MM
approach may be more appropriate.
Acknowledgements This work is partially supported as part of an EPSRC low carbon fuels grant
(EP/N009533/1).
References
1. Finkelstein J (2009) Metalloproteins Nat 460:813–813. https://doi.org/10.1038/460813a
2. Lu Y, Yeung N, Sieracki N, Marshall NM (2009) Design of functional metalloproteins. Nature
460:855–862. https://doi.org/10.1038/nature08304
M. G. Quesne and S. P. de Visser
strate approach for the correct orbital overlap is approximately 180°, which usually
results in less stereochemical clashes than seen in the optimal 120° of the π-pathway
and also leads to the maximum level of exchange interactions between unpaired electrons on Fe(III) [52, 113, 116–121]. The σ -pathway for hydrogen-transfer barrier
(
5 TS H, σ ) included a Fe–O–C angle of 141.4° which is obviously considerable smaller
than the optimal; while the angle of substrate approach seen in
5 TS Hπ (131.4°) is
considerably more favorable for π hydrogen abstraction. The potential energy surfaces produced by both pathways (from 5 R to 5 I) are shown in Fig. 8 and illustrate
that both
5 TS H,σ and
5 TS H,π have comparable energies: 18.2 versus 18.6 kcal mol
−1 ,
which is considerably more realistic then the value of over 20 kcal mol
−1 found using
cluster models.
The final step of the reaction mechanism involves a rebound of the hydroxyl group
back onto the radical substrate intermediate, with both
5 I σ and
5 I π forming strong
hydrogen-bonding interaction between carboxylate group of Asp 133 and the Fe(III)
− OH. This then leads to relatively high rebound barriers of >13.1 kcal mol
−1 ,
which are very far away from the 0.6 kcal mol
−1 found with the DFT cluster model
calculations [115].
5 Conclusion
Computational studies on non-heme and heme monoxygenases and dioxygenases
are useful for the understanding of metalloenzymes and in particularly for providing
insights into the chemical and physical properties of short-lived and highly reactive
intermediates. The key to these studies remains the choice of the model, which should
be a good representation of the real system. Two approaches have been used, namely
QM models and QM/MM techniques. The latter take the structure of the full enzyme
with a solvent layer into consideration whereas the former only include key features
of the first and second coordination sphere. We have shown several examples where
small active site models using the QM cluster technique give viable results that
compare well with experiment. For systems, however, where substrate binding is
tight and or a network of hydrogen-bonding interactions exist, a complete QM/MM
approach may be more appropriate.
Acknowledgements This work is partially supported as part of an EPSRC low carbon fuels grant
(EP/N009533/1).
References
1. Finkelstein J (2009) Metalloproteins Nat 460:813–813. https://doi.org/10.1038/460813a
2. Lu Y, Yeung N, Sieracki N, Marshall NM (2009) Design of functional metalloproteins. Nature
460:855–862. https://doi.org/10.1038/nature08304