Computational Study of Engineered Cytochrome P450-Catalyzed C–H Amination: The Origin of the Regio- and Stereoselectivity

Cytochrome
P450 enzymes were recently engineered to catalyze the
C–H amination reaction of aryl sulfonyl azides with excellent
regio- and stereoselectivity (Arnold and co-workers J. Am. Chem. Soc. 2014, 136, 15505). The mechanism
of this reaction was studied by quantum mechanical (QM)/molecular
mechanical (MM) calculations in this work. The C–H activation
is found to be a stepwise process consisting of hydrogen abstraction
(H-abstraction) of the reactive C–H bond by an iron nitrenoid
cofactor to produce the biradical intermediate and subsequent radical
rebinding to form the final product. The rate of rotation of the carbon
radical center was estimated to be much faster than that of radical
rebinding, which implies that the H-abstraction does not determine
the stereoselectivity. For mutant A, the H-abstraction step has a
barrier of 16.7 kcal/mol, which is 3.0 kcal/mol higher than that of
the following radical rebinding step. The H-abstraction step determines
the regioselectivity, but the radical rebinding step determines the
stereoselectivity. Barriers of these two steps are 16.1 and 27.5 kcal/mol,
respectively, for mutant B. It is different from mutant A in that
the radical rebinding step has the higher barrier and determines both
the regio- and stereoselectivity. The initial distances between the
hydrogens of reactive C–H bonds and the iron nitrenoid were
found to not correlate with their reactivities. The calculated barriers
are qualitatively consistent with the experimentally observed regio-
and stereoselectivity with the exception of the stereoselectivity
of mutant B. The lower barriers of mutant A presumably come from the
stabilization effect of the H-bond between G265 and the sulfone O.
This H-bond does not exist in mutant B. The conformation of the protein
backbone, with the exception of the active site, does not change much
(RMSD < 0.05) along the reaction pathway.