Communications
intermediate Int, which then abstracts a second hydrogen-
tion reactions by a range of iron(IV)–oxo oxidants implicated
atom via the transition state TSD to form anthracene and MnIII
as products (Prod). Calculated free energy values show that
the first hydrogen-atom abstraction barrier (TSHA) is rate-
determining, and is in agreement with the experimental
finding that there is no observable build-up of a MnIV
intermediate. A triplet spin state 3[(Cz)Mn(O)] (spin densities
reveal a [(Cz+C)MnIV(O)] configuration) lies well above the
ground state for the reactants by 6.2–9.5 kcalmolꢀ1, but does
react to form a radical intermediate 3Int with the same orbital
occupation as the singlet; [(Cz)MnIV(OH)]–[DHA-H]C. The
potential energy profiles for the triplet states for 1, [1-F]ꢀ, and
[1-CN]ꢀ (Figures S15–S17 in the Supporting Information)
a linear correlation of barrier height with BDEOH
,
in
[27]
agreement with the trend observed here (Figure S20 in the
Supporting Information). Mayer et al. have shown that the
BDEs and rates of HAT for metal complexes correlate with
simple organic radicals.[21]
A plot of the theoretical
BDEOH values versus experimental log k for the reactions of
1, [1-F]ꢀ, and [1-CN]ꢀ with DHA fits remarkably well with
the data for the organic radicals sec-BuO2C and tBuOC, thereby
enhancing the validity of the BDEs derived from DFT-
(Figure 3).
3
reveal transition states TSHA slightly higher in energy (0.8–
1
1.6 kcalmolꢀ1) than the corresponding singlet states TSHA
.
The singlet and triplet states become nearly degenerate upon
reaching the intermediate, Int, and therefore a spin-state
crossing may occur, but only following the initial hydrogen-
atom abstraction barrier TSHA. Hence, the reactivity of 1, [1-
F]ꢀ, and [1-CN]ꢀ toward HAT are not influenced by a spin-
state crossing mechanism. Optimized geometries of TSHA are
shown in Figure 2. Without an axial ligand, the barrier is late
ꢀ
ꢀ
with short O H and long C H bond lengths, which is typical
for high-barrier transition states.[23]
The lack of multistate reactivity for 1 is in contrast to HAT
reactions involving many other heme and nonheme metal–
oxo species, for which multistate reactivity has been invoked
as a critical factor.[7,24] These results also contrast recent
theoretical work on MnV(O) porphyrins, for which multistate
reactivity has been suggested.[25] Interestingly, the singlet state
of MnV(O) porphyrins was found completely unreactive
toward HAT, but our calculations suggest that the MnV(O)
corrolazines, in contrast, likely operate via a novel single-
state-reactivity pathway.
What is most striking from DFT calculations is the
dramatic lowering of the initial HAT barrier (TSHA) upon
the coordination of Fꢀ or CNꢀ to 1 (7.4 and 9.7 kcalmolꢀ1
lower than 1, respectively). The trend in reactivity, [1-CN]ꢀ >
[1-F]ꢀ @ 1, predicted in Figure 2 by the relative HAT barriers
(TSHA) is in excellent agreement with the experimental rate
constants for DHA oxidation. Calculated KIE values of 8.5
and 8.6 for hydrogen-abstraction by [1-F]ꢀ and [1-CN]ꢀ,
respectively, correspond well with the experimental findings
(Table 1).
Figure 3. Plot of DFT calculated BDEOH values versus experimental log
k2’ values for the reaction of 1, [1-F]ꢀ, and [1-CN]ꢀ with DHA. Best-fit
line through data for sec-BuO2C and tBuOC.[21]
A thermodynamic analysis[21] shows that the strength of
1
IV
ꢀ
the O H bond in [(Cz)Mn (OH)] is proportional to the
electron affinity (EA) of the MnV(O) complex and the proton
affinity of the hypothetical one-electron-reduced [MnIV(O)]ꢀ
species. The EA values for 1, [1-F]ꢀ, and [1-CN]ꢀ obtained by
DFT calculation (Table S14 in the Supporting Information)
reveal a significant decrease (11.3–14.0 kcalmolꢀ1) upon
addition of the anionic axial donors compared to 1, which
ꢀ
should induce a weakening of the O H bond in the
Mn complex. Thus, according to the DFT calculations, the
large increase in BDEOH must arise from a substantial
increase in proton affinity (26.4–28.7 kcalmolꢀ1) upon addi-
tion of Fꢀ or CNꢀ. In other words, the basicity of [MnIV(O)]ꢀ,
not the oxidizing power of the MnV(O) complex, is predicted
to be greatly enhanced by axial ligation and leads to the
dramatic increase in reactivity.[28]
What drives the lowering of the HAT barrier upon
addition of anionic donors? A clue comes from the stability of
These results are in concert with the notion that the Cys
ligand in Cyt-P450 may serve to enhance the basicity of the
iron–oxo intermediate, thereby providing a means for the
the 1Int intermediate, which increases significantly upon
ꢀ
coordination of Fꢀ or CN . The C H activation step for
ꢀ
nonligated 1 is endothermic, whereas the same step for [1-F]ꢀ
and [1-CN]ꢀ is strongly exothermic. Because reaction exo-
thermicity and barrier height often correlate with each
activation of inert C H bonds in the presence of an
ꢀ
oxidatively susceptible protein matrix. Interestingly, recent
DFT calculations on Cyt-P450 suggest that the anionic Cys
donor induces similar opposing effects on electron affinity
and proton affinity.[29] Further studies to understand the
influence of ancillary ligands on metal–oxo HAT reactions in
both synthetic and biological systems are clearly warranted.
other,[26] it is important to determine the possible factors
1
ꢀ
that influence the stability of Int. Calculation of the O H
bond strength for 1[(Cz)MnIV(OH)] gives
a
BDEOH
=
69.3 kcalmolꢀ1, which dramatically increases to 81.7 and
86.7 kcalmolꢀ1 for the systems with Fꢀ or CNꢀ, respectively.
These values predict that the exothermicity of HAT should
increase by 12.4 and 17.4 kcalmolꢀ1, respectively, as a result of
addition of Fꢀ or CNꢀ. Recent studies of hydrogen-abstrac-
Received: February 26, 2010
Revised: April 21, 2010
Published online: June 16, 2010
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2010, 49, 5091 –5095