Alkane Hydroxylation by a Nonheme Iron Catalyst that Challenges the Heme Paradigm for Oxygenase Action

,

RECEIVED DATE (will be automatically inserted after manuscript is accepted) Selective oxidation of hydrocarbons under mild conditions constitutes a major challenge of modern chemistry. 1 Nonheme iron enzymes like methane monooxygenase 2 and Rieske dioxygenases 3 catalyze such reactions and have inspired the development of synthetic models as alkane oxidation catalysts. 4 Exceptional nonheme iron catalysts are those containing tetradentate N4 ligands belonging to the tripodal TPA and linear BPMEN family that can perform efficient stereospecific alkane hydroxylation using H2O2 as oxidant. 4 These bioinspired catalysts have provided key insights into the mechanisms by which alkanes are oxidized by enzymes. The common mechanism invoked in these studies follows the heme paradigm 5 in which a high valent iron-oxo species first abstracts an H-atom from a substrate C-H bond and the incipient alkyl radical then rebounds with the nascent Fe-OH moiety to form the product C-O bond (Scheme 1). 2,4 Labeling studies indicate that the O atom introduced into the substrate C-H bond derives mainly from the H2O2 oxidant. However a minor fraction incorporates an O atom from water into the product, which is explained by oxo-hydroxo tautomerism of the high-valent Fe=O oxidant. 5,6 Herein we describe a new remarkably efficient nonheme iron catalyst in the stereospecific oxidation of alkanes using H2O2 as oxidant that yields alcohol products with an unexpectedly large fraction of their oxygen atoms derived from water. This unprecedented result challenges the established heme paradigm and sheds new light into how nonheme iron centers may differ from heme centers in their mode of oxidative action.
(1) (see Figure 1) was prepared and structurally characterized (for details, see Sup. info). Its crystal structure presents a distorted octahedral iron(II) center having a tetradentate Me2 PyTACN ligand and two triflate anions coordinated cis to each other.
Complex 1 is a very efficient catalyst for the hydroxylation of cyclohexane by H2O2. When 10 equiv H2O2 were delivered by syringe pump together with 1000 equiv water, 6.0 turnover numbers (TN) of cyclohexanol (A) and 0.5 TN of cyclohexanone (K) (A/K= 12) were obtained, corresponding to a 65% efficiency in the conversion of H2O2 into products. Both efficiency and A/K selectivity exhibited by 1 were comparable to values reported for [Fe II (BPMEN)(CH3CN)2] 2+ (63% yield and A/K = 8), considered the prototypical example of an efficient stereospecific alkane hydroxylation catalyst. 4b Additional evidence implicated a metalbased oxidant rather than HO•. The intermolecular kinetic isotope effect evaluated in the hydroxylation of a 1:3 molar mixture of cyclohexane and its deutero-analogue was 4.3. In addition 1 oxidized adamantane with a large 3°/2° normalized selectivity (30). Lastly, the oxidation of cis-1,2-dimethylcyclohexane (cis-DMCH) afforded the tertiary alcohol product with 93% retention of configuration. Overall, these characteristics are consistent with those of some previously described nonheme iron complexes. 4 Interestingly, 18 O-labeling experiments in the hydroxylation of different alkanes catalyzed by 1 revealed intriguing differences between this catalyst and previously reported non-heme N4Fe systems. The oxidation of cyclohexane by 1 with 10 equiv H2 18 O2 in the presence of 1000 equiv H2 16 O afforded 47% 18 O-labeled cyclohexanol. Complementary experiments with 10 equiv H2O2 and 1000 equiv H2 18 O yielded 42(3)% 18 O-labeled cyclohexanol. The same levels of label incorporation from H2 18 O were observed in the hydroxylation of cyclooctane (Table 1). This level of oxygen incorporation from water was the highest for any synthetic nonheme catalyst thus far. Furthermore, when oxidations of cyclohexane were carried out with different H2 18 O concentrations, the fraction of 18 O-labeled cyclohexanol (% R 18 OH) increased linearly with the amount of H2 18 O at lower H2 18 O concentrations, but plateaued at 42(3)% at higher concentrations ( Figure S2), suggesting the involvement of a water binding pre-equilibrium, as precedented for Fe(TPA). 4b These results support the mechanism shown in Scheme 1 proposed for Fe(TPA) and Fe(BPMEN) catalysis in which a HO-Fe V =O oxidant is formed via water-assisted heterolysis of the O-O bond of a H2O-Fe III -OOH intermediate. 4b,7 If complete oxo-hydroxo tautomerism 5,6 of the HO-Fe V =O moiety occurred, a maximum value of 50% would be expected for the amount of labeled water incorporated into the alcohol product. The H2 18 O labeling results described above clearly approach this maximum, so that the C-O bond that is formed can be made with either the oxygen from H2O2 or the oxygen from H2O with almost equal probability. Surprisingly, even higher levels of H2 18 O incorporation (76±3 %) were obtained for alcohol products in the oxidation of alkanes with tertiary C-H bonds, such as adamantane, cis-DMCH, and 2,3-dimethylbutane. Furthermore this unexpectedly large value was found to be independent of substrate concentration (25-1000 mM). This observation strongly implicates the HO-Fe V =O species as the only oxidant capable of alkane oxidation in the case of 1.
The labeling results for 1 differ significantly from those reported for Fe(TPA) ( Table 1). For the latter catalyst, label incorporation from H2 18 O into substrates with 2° C-H bonds decreased with the C-H bond strength and labeling of 2°-ol products was much higher than for 3°-ol products. These results suggested that C-H bond cleavage and oxo-hydroxo tautomerism were competitive processes, as also found for iron porphyrin complexes. 4b,5,6 However this was not the case for 1, since higher water incorporation was observed in the oxidation of the weaker 3° C-H bonds and this level of incorporation was independent of cis-DMCH concentration. To rationalize the high level of label incorporation, we considered the possibility that a carbocation intermediate was formed and subsequently trapped by water, but discarded it because of the high retention of configuration observed for cis-DMCH hydroxylation. We also considered the possibility that the HO-Fe V =O oxidant became doubly labeled by rapid, multiple intermolecular exchanges with H2 18 O, but rejected it on the basis of a H2 18 O labeling experiment where cyclohexane and cis-DMCH were oxidized in competition with each other. cis-DMCH reacted much faster than cyclohexane, and so should have a shorter lived oxidant. Nevertheless the 3 o -ol thus formed contained 74% label from water while cyclohexanol contained only 38%. Thus the difference in label incorporation cannot be determined by how much 18 O is present in the common Fe V oxidant, and a modified mechanistic scenario is required for 1.
In Scheme 2, we propose a scenario in which the structure of the substrate can play a role in determining the course of C-O bond formation. For 1, the two Fe V (O)(OH) isomers (A and B) in equilibrium via oxo-hydroxo tautomerism are distinct, since the ligands trans to the oxo and hydroxo groups are not chemically equivalent and the orientations of the pyridine ligand relative to the Fe=O bond are different. No matter which isomer abstracts the H-atom from substrate, a common Fe IV (OH)2 species is formed. The cis configuration of the two hydroxo groups differs from the trans configuration required in heme complexes and is a unique feature of these nonheme iron catalysts. Thus the alkyl radical can in principle rebound with either OH group, the outcome of which determines the %H2 18 O incorporation. Given that these values range from 40% for cyclohexane-d12 to 79% for cis-DMCH, the energy difference between the two possible rebound trajectories is < 2 kcal mol -1 , which is consistent with DFT calculations. 8 Thus the labeling results indicate that 2°-alkyl radicals do not discriminate between the two OH groups, but 3°alkyl radicals favor rebound with the OH group derived from water. Perusal of the structure of 1 (Figure 1) suggests steric effects may provide a possible rationale for this preference, since the hemisphere surrounding O1 is less sterically congested.

Scheme 2
In conclusion, 1 is an efficient alkane hydroxylation catalyst that can incorporate surprisingly large amounts of water into products via an unusual rebound-like mechanism. A similar mechanism may be operating in the hydroxylation of indane to 1indanol by toluene or naphthalene dioxygenase to account for the high level of H2 18 O incorporation (68%) found. 9 In addition, the α-ketoglutarate-dependent halogenases represent an established biological example in which C-H bond cleavage initiated by an Fe=O moiety is consummated by rebound not with the incipient OH group but with the adjacent halide ligand. 10 Thus 1 may serve as a model for this unusual enzymatic chemistry, which suggests that a broader mechanistic landscape than in heme systems applies to non-heme sites. Abstract A nonheme iron catalyst catalyzed stereoselective oxidation of alkanes with H2O2 with remarkable efficiency and exhibiting an unprecedented high incorporation of water into the oxidized products. The present results challenge the canonical description of oxygenases, the standard oxo-hydroxo tautomerism that applies to heme systems and serves as a precedent for alternative pathways for the oxidation of hydrocarbons at nonheme iron oxygenases.