A novel iripodal ligand containing three different N-heterocyclic donor functions and its application in catechol dioxygenase mimicking

Article abstract of DOI:10.1002/chem.200802591

We describe a novel chiral ligand, L, in which three different Ndonor functions are linked to a methoxymethine unit: a methylpyrazole derivative, a methylimidazole unit, and a pyridyl residue. Complexes with FeCl₂, FeBr₂, and FeCl

Full text of DOI:10.1002/chem.200802591

FULL PAPER
DOI: 10.1002/chem.200802591
A Novel Tripodal Ligand Containing Three Different N-Heterocyclic Donor
Functions and Its Application in Catechol Dioxygenase Mimicking
Marit Wagner, Christian Limberg,* and Thomas Tietz[a]
Dedicated to Professor Joachim Sauer on the occasion of his 60th birthday
Abstract: We describe a novel chiral
ligand, L, in which three different N-
di-tert-butylcatechol, triethylamine, and
O2. Both complexes yielded similar re-
sults in such investigations, since the
LFeII–catecholate complex reacts with
O2 through one-electron oxidation in
the first step. Employing 3 in acetoni-
trile solution, intradiol cleavage oc-
curred, although the undesired quinone
was formed as the main product. If re-
agents were added (NaBPh4, H+) or
reaction conditions were chosen
(CH2Cl2 instead of CH3CN as the sol-
vent) that made the coordination
sphere at the iron centre more accessi-
ble for a third substrate donor function,
an alternative reaction route, presuma-
bly involving O2 binding at the metal,
became more important, which led to
extradiol cleavage. In the extreme case
(CH2Cl2 as the solvent and with the ad-
dition of NaBPh4), mainly the extradiol
cleavage products were formed; the in-
tradiol products were only observed as
side products then and quinone forma-
tion became negligible. Protonated
base functions in the second coordina-
tion sphere increased the efficiency of
extradiol cleavage only slightly. The
obtained results are in line with current
understanding of the function of intra-
diol/extradiol dioxygenases.
donor functions are linked to
a
methoxymethine unit: a methylpyra-
zole derivative, a methylimidazole unit,
and a pyridyl residue. Complexes with
FeCl2, FeBr2, and FeCl3 have been syn-
thesized and fully characterized, in-
cluding with respect to their molecular
structures. While in combination with
FeCl3 L coordinates in a tripodal fash-
ion, with FeX2 (X=Cl, Br) it binds
only through two functions and the
pyridyl unit remains dangling. For po-
tential modelling of intradiol and extra-
diol catechol dioxygenase reactivity,
the complexes [LFeCl2], 1, and
[LFeCl3], 3, have been treated with 3,5-
Keywords: coordination modes
·
ꢀ
iron · N ligands · O O activation ·
reaction mechanisms
Introduction
frequently used in this context, the best-known representa-
tive being tetrapodal TPA (tris(pyridylmethyl)amine), in
which three pyridyl donors are linked to an amine.[7] The re-
action pockets of metalloenzymes are intrinsically chiral,
and although certain protein functions do not utilize this
feature directly during substrate conversion, it allows for
stereoselective reactions, for instance at the active centres of
cytochrome P450[8] or the lipoxygenases.[9] We have there-
fore contemplated the synthesis of a chiral tripodal ligand
that contains the donor functions central to biomimetic or
bioinspired chemistry, namely imidazolyl, pyrazolyl, and
pyridyl residues. Here, we report such a ligand, which con-
tains one of each of these units, as well as its iron(II) and
Histidine moieties are often found as ligands within the
prosthetic groups of metalloenzymes.[1–3] A common binding
motif is the so-called 2-histidine-1-carboxylate facial triad,[4]
and in proteins such as tyrosinase,[1] carboanhydrase,[2] or
hemeerythrin[3] as many as three histidine-based imidazole
residues are coordinated to a single metal centre. This has
inspired the design and employment of multipodal ligands
with N-heterocyclic donor functions for biomimetic chemis-
try. For instance, Tp (tris(pyrazolyl)borate)[5] makes use of
three pyrazole donors, while three imidazole units are found
in tris(imidazolyl)methane.[6] Pyridyl-based ligands are also
ironACTHNUGRTENUNG(III) chlorido complexes. Furthermore, we have investi-
gated the reactivity of these compounds in O2 activation re-
actions, and in order to get an idea concerning the potential
of our novel ligand we have chosen a modelling target, in
relation to which various ligands have already been em-
ployed (albeit it does not take advantage of the chirality of
our ligand), namely the extradiol catechol dioxygenase (be-
longing to the 2-histidine-1-carboxylate family).[9]
[a] Dipl.-Chem. M. Wagner, Prof. Dr. C. Limberg, Dipl.-Chem. T. Tietz
Institut fꢀr Anorganische Chemie, Humboldt-Universitꢁt zu Berlin
Brook-Taylor-Str. 2, 12489 Berlin (Germany)
Fax : (+49)030-2093-6966
E-mail: christian.limberg@chemie.hu-berlin.de
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.200802591.
Chem. Eur. J. 2009, 15, 5567 – 5576
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
5567

Results and Discussion
Ligand synthesis: The ligand synthesis started from the
ketone I (see Scheme 1), which has previously been pre-
Scheme 1. Synthesis of the ligand L.
Figure 1. Molecular structure of [LFeCl2] 1. Hydrogen atoms are omitted
for clarity. Selected bond lengths [ꢃ] and angles [8]: Cl1-Fe 2.2607(10),
ꢀ
ꢀ
ꢀ
Cl2 Fe 2.2409(11), Fe N1 2.062(2), Fe N3 2.081(2); N1-Fe-N3 86.06(9),
N1-Fe-Cl2 107.59(7), N3-Fe-Cl2 116.02(8), N1-Fe-Cl1 117.74(7), N3-Fe-
Cl1 110.58(8), Cl2-Fe-Cl1 115.57(5).
pared by Canty et al.[10] and already contains a pyridyl as
well as an imidazolyl moiety. In order to attach a pyrazolyl
unit, a suitable precursor, 3(5)-tert-butylpyrazole,[11] was first
treated with formaldehyde to protect the NH unit.[12] The
protected pyrazole was then deprotonated with 2 equiv of
nBuLi and subsequent reaction with I yielded the alcohol II.
As it was envisaged that the alcohol unit might disturb coor-
dination through the three N-donor functions, II was further
reacted with NaH and MeI in order to methylate all acidic
functions. In this way, the potential ligand L containing
three different N-donor functions connected to a methoxy-
methine unit was obtained. The central carbon atom in L is
thus chiral, and the compound was obtained in the form of a
racemic mixture. Analytical HPLC investigations showed
that the enantiomers could be separated by employing a
chiral AD-H column (mobile phase: hexane/isopropanol/
triethylamine, 90:10:0.2), which, however, was not necessary
for the present application concerning catechol oxygenation;
this does not focus on the chirality, and so the mixture of
isomers has been employed in all of the described studies.
any further electronic saturation by coordination of a fifth
donor function. However, it is not obvious why the pyridyl
arm remains dangling as opposed to the pyrazole donor
function. Hitchman et al. performed spectroscopic investiga-
tions of the metal binding at the tripodal ligand (pyrazolyl)2-
AHCTUNGTREG(NNUN pyridyl)CH and found that, in agreement with previous
studies on the analogous monodentate ligands,[14a] the pyri-
dine group in the tripodal ligand produces a slightly stronger
interaction than either of the pyrazole groups.[14b] A favoura-
ble binding of the pyridine function of L would also be ex-
pected considering the pKa values. Assuming that the
degree of back-bonding is not very high (which is reasonable
for FeII) and that all three donor functions have the same
freedom to bind (with respect to steric constraints), the dis-
cussion may be focused on the free electron pairs interacting
with the metal centre, which in turn might be comparable
with H+; in this case, the binding ability should correlate
with the pKa values, and that of pyridine (5.6) lies between
those of 1-methylpyrazole (2.1) and 1-methylimidazole
(7.2).[15] Hence, on this basis, the imidazole function should
be the strongest donor and the pyridyl residue should bind
ahead of the pyrazole entity. Against this background, the
preferred binding of pyrazole can only be rationalized with
reference to steric considerations. Assuming that, as the
strongest donor, imidazole coordinates first, it would seem
that a chelate with an additional s-donor achieves a more
favourable orbital interaction (aligned with the M–N vec-
tors) if this donor is incorporated into the five-membered
ring of pyrazole rather than the six-membered pyridyl ring.
The angles at the iron centre of 1 vary between 86.06(9)8
(N1-Fe-N3) and 117.75(7)8 (N1-Fe-Cl1), indicating a signifi-
cant distortion of the tetrahedral coordination sphere. The
six-membered chelate ring is almost planar. Analogously to
Complex synthesis: Treatment of L with FeCl2 in THF led
to the formation of the complex [LFeCl2], 1 (Scheme 2),
which has been characterized by IR spectroscopy, elemental
analysis, and single-crystal X-ray diffraction analysis.
As found in the case of the other compounds containing
L described below, crystals of 1 contained both enantiomers
in the unit cell. Only the molecular structure of 1 with L in
the (S)-configuration is shown in Figure 1.
Despite its three donor functions, the ligand coordinates
in a bidentate fashion through the imidazolyl and pyrazolyl
donors, while the pyridyl unit remains dangling. Iron(II) is
known to form tetrahedral complexes quite often,[13] and ap-
parently also in this case the iron centre does not require
the synthesis of 1,
a corresponding bromide complex
[LFeBr2], 2, could be prepared in 47% yield starting from
FeBr2 (Scheme 2). Single crystals could also be obtained for
2 by evaporating the solvent from a solution in THF, so that
an X-ray diffraction study could be performed. All bond
lengths and angles (see the Supporting Information, Fig-
ure S1) were found to be very similar to those in 1.
Scheme 2. Synthesis of the iron(II) complexes 1 and 2.
5568
www.chemeurj.org
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 5567 – 5576

Biomimetic Oxidation
FULL PAPER
Next, we turned our attention to iron
(III). Addition of L
case of EDO one glutamate and two histidine units are
bound to FeII (Scheme 4). Hence, after binding of the sub-
strate catechol, their reaction mechanisms in the presence of
O2 are different, as are the products. IDO may be regarded
to a solution of FeCl3 in THF led to an orange-yellow sus-
pension, from which [LFeCl3], 3, could be isolated in 68%
yield (Scheme 3).
Scheme 3. Synthesis of the ironACTHNUTRGNE(NUG III) complex 3.
Complex 3 has been extensively characterized, including
by single-crystal X-ray diffraction analysis. The molecular
structure is shown in Figure 2. As anticipated, L now coordi-
Scheme 4. Proposed reaction mechanisms for intradiol catechol dioxyge-
nase (IDO) and extradiol catechol oxygenase (EDO).
Figure 2. Molecular structure of [LFeCl3] 3. Hydrogen atoms are omitted
ꢀ
for clarity. Selected bond lengths [ꢃ] and angles [8]: Cl1 Fe 2.3009(12),
ꢀ
ꢀ
ꢀ
ꢀ
Cl2 Fe 2.2795(13), Cl3 Fe 2.2962(12), Fe N3 2.146(4), Fe N1 2.172(3),
ꢀ
as an FeII(semiquinone) complex (rather than an FeIII-
AHCTUNGTREG(NNUN catechol) substrate complex), with substantial radical char-
acter at its carbonyl C atoms, which are thus prone to attack
by the biradical O2, thereby yielding an anhydride. In con-
trast, the first step of the EDO system involves the binding
and activation of O2 at the FeII centre, which subsequently
attacks the substrate to generate a lactone.[16]
Against this background, studies aimed at modelling IDO
should strictly speaking should start from an FeIII–catecho-
late complex, while the mimicking of EDO requires FeII–
catecholate units. The few examples of crystallographically
characterized FeII–catecholate complexes contain tetraden-
tate ligands such as TPA, which, together with the bidentate
substrate, occupy six coordination sites at the iron centre.[17]
Hence, there is no free binding site for O2 coordination and
consequently catechol oxidation by O2 cannot proceed in a
biomimetic fashion. FeII is most probably first oxidized to
FeIII so that the subsequent oxidation process resembles that
proposed for IDO; accordingly, the products of such reac-
tions are often those expected for intradiol cleavage. The in-
itial conversion of parent FeII models to FeIII analogues can,
however, also occur for FeII complexes containing tridentate
ligands, which then also exhibit IDO beside EDO reactivity
(due to insufficient site isolation being provided by the
model ligand set).[18]
Fe N5 2.321(3); N3-Fe-N1 77.79(13), N1-Fe-Cl2 90.51(10), N3-Fe-Cl3
91.79(10), Cl2-Fe-Cl3 97.43(5), N3-Fe-Cl1 90.03(10), N1-Fe-Cl1 97.14(9),
Cl2-Fe-Cl1 100.81(5), Cl3-Fe-Cl1 95.25(5), N3-Fe-N5 81.73(12), N1-Fe-
N5 80.06(12), Cl2-Fe-N5 87.09(9), Cl3-Fe-N5 86.20(9), N3-Fe-Cl2
165.01(10), N1-Fe-Cl3 163.77(10), Cl1-Fe-N5 171.68(9).
nates to the iron centre through its three donor functions.
Accordingly, together with the three chlorido ligands a dis-
torted octahedral coordination sphere results, with angles at
the iron centre ranging from 77.79(13)8 (N3-Fe-N1) to
100.81(5)8 (Cl2-Fe-Cl1). The bond between the iron atom
and the imidazole nitrogen atom (Fe N3) is the shortest
(2.146(4) ꢃ) of the iron–nitrogen bonds, while the distance
of the iron centre to the nitrogen atom of the pyridyl resi-
ꢀ
ꢀ
due is the longest (Fe N5 2.321(3) ꢃ).
Reactivity studies: As indicated in the introduction, the syn-
thesized iron complexes have been tested with regard to
their potential to model catechol dioxygenase reactivity,
bearing in mind a comparison of L with other ligands em-
ployed in biomimetic chemistry. There are two classes of
catechol dioxygenases: the extradiol catechol dioxygenases
(EDO) and the intradiol catechol dioxygenases (IDO).[9]
Both contain three amino acids bound at an iron centre; in
the case of IDO one tyrosine and two histidine residues are
coordinated to FeIII (a further tyrosine residue has been
shown to dissociate upon catechol binding),[9] while in the
Various ligands have been employed for the preparation
of [LFeIII(catecholate)] complexes as functional models for
Chem. Eur. J. 2009, 15, 5567 – 5576
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
5569

C. Limberg et al.
IDO (derivatives of catechol such as 3,5-di-tert-butylcate-
chol, DBCH2, are often employed), and some of these com-
plexes have been well characterized.[19,20] In reactions with
O2, such complexes can lead to various products, the most
undesirable (but often unavoidable) of which is the auto-ox-
idation product quinone Q.
of the present system, and so all further experiments were
carried out by starting directly from 3.
Treating 3 with DBCH2 and two equivalents of triethyla-
mine in acetonitrile led to a colour change to purple, indi-
cating the formation of a complex [LFeACTHNUTRGNEUNG(dbc)Cl] (Scheme 6).
The product expected for IDO reactivity would be A (3,5-
di-tert-butyl-1-oxacyclohepta-3,5-diene-2,7-dione; Scheme 5)
or its hydrolysis product 3,5-di-tert-butyl-5-(carboxymethyl)-
2-furanone. Most IDO models with tridentate ligands tend
to generate A along with the extradiol cleavage products B
(4,6-di-tert-butyl-2-pyrone) and
C
(3,5-di-tert-butyl-2-
pyrone), which are obtained as the main products (they are
generated through the EDO pathway followed by CO elimi-
nation).[19] To explain this, it has been suggested that the
alkyl peroxide intermediate formed from such FeIII–IDO
models can decompose via two pathways, and that subtle in-
fluences determine the ultimate product selectivity (see
Scheme 6. Reaction of 3 with 3,5-di-tert-butylcatechol and triethylamine.
below). Only [LFeIII
(dbc)] complexes, in which L is a tetra-
As yet, we have not been able to isolate and fully charac-
terize this complex, but the colour and UV/Vis spectrum in
acetonitrile (the bands and ex-
dentate ligand, selectively lead to A.[20]
tinction coefficients quoted
above are comparable to those
of other DBC complexes[9]) as
well as
a
peak at m/z=
615.2860 in the high-resolution
ESI mass spectrum, which can
be assigned to the ion [LFe-
Scheme 5. Products of the oxidative cleavage of 3,5-di-tert-butylcatechol.
(dbc)]+ (m/z 615.2866), give
ACHTUNGTRENNUNG
good indications of its forma-
tion. Due to the three different
Complex 1 contains an iron(II) centre bound by L, and
after further coordination by DBCHꢀ the resulting complex
might thus be regarded as an EDO model. Hence, to simu-
late the situation shown on the right-hand site of Scheme 4
we first treated 1 with DBCH2 in the presence of one equiv-
alent of NEt3, which led to a vinaceous wine-red solution.
Attempts to follow the reaction by means of NMR spectros-
copy did not, however, lead to any further information due
to the large half widths of the paramagnetically shifted sig-
nals leading to substantial overlap. Addition of O2 to such a
solution led to an immediate colour change to purple, which
was monitored by UV/Vis spectrophotometry. After 20 min,
new bands had appeared at around 545 nm (e=
1480mꢀ1 cmꢀ1) and 837 nm (e=1560mꢀ1 cmꢀ1), which were
similar to those observed for the independently synthesized
N donor functions of L, three isomers can be expected upon
binding of a chelating ligand to the [LFeCl]2+ unit, and due
to the asymmetric substitution pattern within DBC a further
three are possible. Preliminary calculations (BP86/
LANL2DZ) showed that the energies of the individual iso-
mers vary within a range of 12.2 kJmolꢀ1 (see the Support-
ing Information, Table S2). Attempts to crystallize the prod-
uct from such solutions led in one case to a small amount of
crystals of 1 (slow evaporation of the solvent from a solution
in diethyl ether of the product obtained from a reaction per-
formed in dichloro
ACHTUNGTRENNUNG
of [Cl2Fe(m-dbc)2FeCl2]ACHTNUGTRENNNUG
N
solvent after performing the reaction in acetonitrile;
Figure 3). Comparison of the structural data of the [FeL2]
ꢀ
ꢀ
ꢀ
moiety of 4 (Fe NPz 1.990(5) ꢃ, Fe NIm 2.007(5) ꢃ, Fe NPy
[LFeIII
e=1347mꢀ1 cmꢀ1
ure S2). The latter two bands could be assigned to catecho-
lato–iron(III) LMCT transitions, which are characteristic of
G
1.985(5) ꢃ) with those of the complex cation [L2Fe]2+ of the
;
independently synthesized iron(II) compound [FeL2ACHTUNGTRENNUNG(OTf)2]
ꢀ
ꢀ
ꢀ
(5) (Fe NPz 2.1093(19) ꢃ, Fe NIm 2.0797(19) ꢃ, Fe NPy
2.194(2) ꢃ; see the Supporting Information; Figure S4), as
well as with those of the LFe3+ unit in 3, points to an oxida-
ACHTUNGTRENNUNG
such moieties.[21] Such rapid FeII-to-FeIII–catecholato oxida-
tions upon exposure to O2 have been observed before (see
above) and as a result of this transformation the reactivity
found for LFeII/DBC/O2 systems will clearly be very similar
to that observed for the corresponding LFeIII/DBC/O2 sys-
tems.[18] Preliminary experiments (see the Supporting Infor-
mation; Table S1) showed that this was also true in the case
tion state of +II within the cation of 4, so that the [Cl2FeACHTUNGTRENNUNG(m-
dbc)2FeCl2] moiety should bear a double negative charge.
This is consistent with the observation that the compound
(lutidinium)2ACTHUNTRGENNUG[Cl2FeACHTUNTGREN(NNGU m-dbc)2FeCl2], with a similar anion, has
been obtained from the reaction of FeCl3 with 3,5-di-tert-bu-
tylcatechol and two equivalents of 2,6-lutidine in THF.[22]
5570
www.chemeurj.org
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 5567 – 5576

Biomimetic Oxidation
FULL PAPER
Studies concerning the reac-
tivity of [LFeACTHNUTRGENUGN(dbc)Cl] in the
presence of O2 were started as
follows: 3, DBCH2, and NEt3
in a ratio of 1:1:2 were dis-
solved in acetonitrile, excess
O2 was introduced, and the re-
action was monitored by UV/
Vis
(Figure 4). Both bands in the
visible region (554 and
spectrophotometry
813 nm) decreased during the
reaction and a new band devel-
oped at 379 nm, while an iso-
sbestic point could be observed
at 441 nm. As the concentra-
tion of dioxygen in the acetoni-
trile (8.1 mm)[20f] was much
higher than that of the reacting
complex (0.3 mm), the reaction
conformed to pseudo-first-
order kinetics with a rate con-
stant of k(O2)=0.0123mꢀ1 sꢀ1
(Figure 4). Within a reaction
time of six hours, the original
Figure 3. Molecular structure of [Cl2FeACHTNUTRGENNG(U m-dbc)2FeCl2]CAHTUNGTERN[NUNG FeL2], 4. Hydrogen atoms are omitted for clarity. Select-
ꢀ
ꢀ
ꢀ
ꢀ
ed bond lengths [ꢃ] and angles [8]: Cl1 Fe1 2.2530(11), Cl2 Fe1 2.2691(12), Fe1 O1 1.909(3), Fe1 O2
ꢀ
ꢀ
ꢀ
ꢀ
2.030(3), Fe1 O2’ 2.011(2), Fe2 N1 2.000(3), Fe2 N3 2.008(3), Fe2 N5 1.974(3); O1-Fe1-O2 79.63(10), O1-
Fe1-O2’ 149.19(11), O2-Fe1-O2’ 70.95(11), O1-Fe1-Cl1 103.58(9), O2-Fe1-Cl1 116.69(9), O2’-Fe1-Cl1 97.73(8),
O1-Fe1-Cl2 95.89(9), O2-Fe1-Cl2 133.48(9), O2’-Fe1-Cl2 97.72(8), Cl1-Fe1-Cl2 109.41(5), N5-Fe2-N1
93.97(13), N5-Fe2-N1’ 86.03(13), N5-Fe2-N3 88.43(14), N5-Fe2-N3’ 91.57(14), N1-Fe2-N3 92.62(14), N1-Fe2-
N3’ 87.38(13).
LMCT bands of the ironACHTUNGTRENNUNG(III)–
In the anion of 4 the two iron centres are each coordinat-
ed by two chlorido ligands, and they are bridged by two bi-
dentate catecholato units, whereas in (lutidinium)2ACTHNUTRGNENUG[Cl2FeACHTUNGTERN(NUGN m-
dbc)2FeCl2] the catecholate moieties bind only through one
oxygen donor.
catecholato complex had disappeared almost completely.
The rate constants found for other relevant systems,[20c,e,f,21]
vary over three orders of magnitude and that found here
falls at the lower end of the range. While one might expect
to find a correlation between the CT band position and the
reactivity, this is not evident: the rate constant measured
here is significantly different from those found for com-
plexes with similar CT band positions, and these in turn
show great variations in reactivity among themselves.
Although 1 and 4 were only isolated in small quantities as
side-products, their formation shows that [LFeACTHNUTRGENUG(N dbc)Cl] is
quite labile and slowly decomposes through redox reactions
to give FeII compounds. This is not surprising as IDO reac-
tivity is based on a catecholato-to-iron
(III) electron transfer
For an analysis of the reaction products, all volatiles were
removed from the reaction mixture after 6 h and the residue
(see above). The degradation of 1:1 FeIII–catecholate com-
plexes to produce FeII has been reported previously, and as
an acceleration by light was mentioned in this context,[23]
the influence of light was also considered in the subsequent
investigation of O2 reactivity here. However, light did not
seem to play a major role in our case (see below). For the
reactivity studies with O2 described below, it was assumed
that the extent of decomposition during the reaction time
was small, so that the concentrations of 1, 4, or other prod-
ucts of degradation could be regarded as negligible. This is
important, since reactions of anions such as [Cl2FeACTHNUTRGENUG(N m-
dbc)2FeCl2]2ꢀ would surely lead mainly to Q; indeed, a test
experiment showed that FeCl3 reacted with DBC2ꢀ and O2
in acetonitrile within six hours to give almost exclusively Q
(99%, accompanied by 1% B), which is in agreement with
previous investigations.[24] ESI-MS studies confirmed the val-
idity of the above assumption: upon addition of DBCH2 and
NEt3 to a solution of 3 in acetonitrile, a peak at m/z
615.2860 immediately appeared and storage of the solution
for 6 h did not lead to any further peaks.
Figure 4. Evolution of the absorption spectrum upon exposure of 3 +
DBCH2 + 2NEt3 to O2 in MeCN. Spectra were recorded at hourly inter-
vals. The top line shows the absorption spectrum of 3 + DBCH2
2NEt3. Inset: plot of lnc vs time.
+
Chem. Eur. J. 2009, 15, 5567 – 5576
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
5571

C. Limberg et al.
Table 1. Oxidative catechol cleavage reactions with O2.[a]
three days are not much higher
than those obtained after six
hours. For further investiga-
tions, it thus seemed reasona-
ble to continue comparing re-
sults obtained after 6 h, like
those presented in Table 1,
where entry 3 corresponds to
the fourth measurement in
Figure 5. It shows that the ad-
Entry
Solvent
LFeCl3 (3)
Additive
Q [%]
A [%]
B [%]
C [%]
B+C [%]
1
2
3
4
5
6
MeCN
MeCN
MeCN
CH2Cl2
MeCN
CH2Cl2
DBC2ꢀ
DBC2ꢀ
DBC2ꢀ
DBC2ꢀ
DBCHꢀ
DBCHꢀ
–
–
63
65
61
12
58
6
34
32
31
22
20
23
3
3
8
42
13
43
–
–
3
3
8
66
22
71
[b]
NaBPh4
NaBPh4
NaBPh4
NaBPh4
trace
24
9
28
[a] After 6 h reaction time, conversions for all reactions are between 77 and 85%. [b] The reaction was per-
formed in the absence of light.
was extracted with CHCl3. After filtration of the extract
through silica gel, product yields were determined with the
aid of 1H NMR and GC-MS/FID (entry 1 in Table 1). It
turned out that, under these conditions, mainly Q was
formed, together with half as much A, the formation of
which as a main product was not unexpected regarding 3 as
an IDO model. Indeed, the yield of extradiol products was
very low. Referring to the apparent instability of 3 in at-
tempts to crystallize it, entry 2 describes an experiment in
which the influence of light was tested (see above).[23] How-
ever, entries 1 and 2 are nearly identical, suggesting that
light does not have any effect. Entry 3 in Table 1 addresses
the influence of the coordination number at the iron centre,
since the availability of a vacant site has been argued to be
an essential factor for the occurrence of extradiol-type cate-
chol cleavage:[16] O2 then has the additional possibility of co-
ordinating to the iron(II) centre generated in the course of
electron transfer from the DBC ligand. Subsequently, an in-
termediate is formed, in which the organoperoxide unit is
Figure 5. Temporal dependence of the product ratios of the reaction be-
tween 3/DBC2ꢀ/NaBPh4 and O2 in MeCN.
dition of NaBPh4 does indeed somewhat decrease the yields
of A and Q in favour of B; however, the effect is not very
large, probably because acetonitrile solvent molecules
occupy the free coordination sites. Hence, the same reaction
was repeated in dichloromethane (entry 4), which led to dra-
matic changes: the yield of Q was significantly decreased,
while the dioxygenase reactivity was changed from mainly
intradiol to mainly extradiol. This is plausible since dichloro-
coordinated through three donor functions to an ironACHTUNTRGNEUNG(III)
centre (Scheme 7), in a fashion similar to that proposed for
the EDO intermediate (cf. Scheme 4), in which, however,
the iron atom is in the +II oxidation state.[16]
Accordingly, NaBPh4 was added, with the intention of ab-
stracting the residual chlorido ligand. In order to ensure that
the appropriate reaction times were chosen, the reaction
course was followed with time (aliquots were removed and
analyzed). The results are shown in Figure 5.
The results obtained for Q and DBCH2 are somewhat un-
reliable during the first hours of the conversion, as unreact-
ed DBCH2 is partly converted into Q during work-up. It is
therefore more informative to concentrate on the other
products. Evidently, the concentrations of A and B increase
with increasing reaction time, but the yields obtained after
AHCTUNGTREGmNNUN ethane cannot occupy the free coordination site effective-
ly, and apparently, as previously observed in cases of facially
coordinating tridentate ligands,[19] this leads to a change in
mechanism with respect to O2 reactivity, as outlined above.
Finally, the influence of the presence of protons was investi-
gated for two reasons: first, it has been noted in investiga-
tions of models that the presence of protons may lead to the
protonation of donor functions and thus to more spacious
coordination spheres.[25] Secondly, it has been found in theo-
retical and experimental inves-
tigations concerning the native
and mutated enzymes that the
presence of a protonated histi-
dine residue in the second co-
ordination sphere is decisive
for extradiol vs intradiol cleav-
age because it is thought to
stabilize the iron–superoxo
species,[26] to accelerate the
attack of the bound superoxide
Scheme 7. Proposed mechanism for the reaction of 3 in the presence of DBCH2, 2 NEt3, NaBPh4, and O2.
on the substrate,[27] or to pro-
5572
www.chemeurj.org
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 5567 – 5576

Biomimetic Oxidation
FULL PAPER
mote alkenyl migration in the peroxo intermediate yielding
the lactone species.[28] We have addressed this issue by
studying the effect of reducing the amount of NEt3 added to
only one equivalent. Comparison of entries 1 and 5 shows
that this somewhat decreased the yield of A, while more B
and even some C were formed. A small increase in extradiol
reactivity due to the presence of protons was also observed
for solutions in dichloromethane (cf. entries 4 and 6).
Hence, the effect of the presence of protons is greater than
that of generating a free coordination site in acetonitrile,
but smaller than that of a vacant site in dichloromethane.
Assuming that, for the reasons given above, there is a cor-
relation between the space available at the metal centre and
the extent of extradiol reactivity, these results can be ration-
alized as follows. In acetonitrile, a vacant site is immediately
occupied by this solvent, so that hardly any difference is
noted in comparison to chlorido coordination. The space
generated by the free site is only available in dichlorome-
thane, and this leads to significantly increased EDO reactivi-
ty. Addition of a proton (or more precisely omission of a
base equivalent so that one of the DBCH2 protons remains
within the system) can be expected to have the two effects
outlined above: (i) protonation of one ligand donor function,
thereby generating a free coordination site, and (ii) an influ-
ence on the superoxide or peroxide intermediates that pro-
mote extradiol cleavage. A comparison of entries 4 and 6
allows evaluation of point (ii): the conditions of entry 4 are
hardly any formation of Q is observed and the extradiol re-
activity reaches the highest value of 71%.
Intradiol cleavage seems to require only two coordination
sites (for the coordination of DBC2ꢀ) and was thus observed
in all of the experiments.[30] The extent was highest (34%)
when extradiol reactivity was basically excluded and some-
what smaller (20–23%) with competing extradiol cleavage.
If extradiol reactivity is excluded through the lack of a free
coordination site, only the formation of Q occurs simultane-
ously. With increasing available space, however, extradiol
cleavage becomes more efficient than auto-oxidation, so
that Q formation decreases in favour of extradiol cleavage.
Altogether, the results can be understood on the basis of
previous hypotheses made by considering the reactivities of
model compounds and native enzymes. Although there is a
different order of initial events in the intradiol vs extradiol
reaction mechanisms, they converge on a common proximal
organoperoxide intermediate.[16] The two enzymes differ in
the oxidation state of the iron centre within this intermedi-
ate, but considering the facts that the IDO pyrocatechase
can perform extradiol cleavage when 3-methylcatechol is
employed as the substrate[31] and that following mutation of
certain EDOs (replacement of a His residue in the active
site by Phe) intradiol cleavage is observed,[32] the choice of
intradiol vs extradiol reaction pathways is clearly deter-
mined not so much by the iron oxidation states, but rather
by stereoelectronic factors influencing the acyl vs alkenyl
migration rearrangements of the organoperoxide intermedi-
ates.[16] The extradiol dioxygenase active site containing the
His2Glu motif is able to bind substrates and reaction inter-
mediates through three coordination sites, and hence the or-
ganoperoxide intermediate may also be bound in a triden-
tate coordination mode (compare Scheme 4), that is, in a
conformation in which the peroxide unit is positioned pseu-
doaxially.[16] This alignment allows for facial alkenyl migra-
tion and extradiol cleavage. The same can be deduced for
iron(II) models containing facial tridentate ligands, as their
coordination environment resembles the situation in the re-
spective dioxygenase active site, and the conclusion even
expected to produce [LFeACHTNUTGRNEUNG
(dbc)]+ with a free coordination
site, while those of entry 6 are likely to also produce a pen-
tacoordinated iron centre, with the difference that L is only
coordinated through two donor functions leaving one pro-
tonated base dangling; a chlorido ligand is left to coordinate
instead. Hence, the situations in entries 4 and 6 differ in the
availability of a protonated base function in the second co-
ordination sphere, and this leads to an increase of the yields
of extradiol products by 5%. Accordingly, the presence of
BH+ (B=base) has only a small effect. Applying the same
arguments, entry 5 may be compared with entry 3, but a
much larger influence (14%) is observed in this case, so that
a further aspect must also play a role. While in the case of
entry 3 acetonitrile can be expected to occupy the free coor-
dination site generated, this might not be possible with the
same efficiency in the case of entry 5 as the BH+ remains in
close proximity to the metal centre and blocks the linear co-
ordination of acetonitrile to some extent (similar to the way
in which a dangling His residue in hemoglobin disturbs the
coordination of CO vs O2).[29] Accordingly, a third coordina-
tion site is more readily available (albeit not completely as
in the case of entry 4 since there is still competition with
acetonitrile), and this increases the yield of extradiol prod-
ucts. If this hypothesis is correct, the space available at the
iron centre systematically increases in the series of entries 1
! 3 ! 5 ! 4, and indeed the degree of extradiol reactivity
increases in the same order. Entry 6 can be considered as
the extreme case, since under these conditions the iron
centre can be expected to have a free coordination site in
addition to the beneficial influence of BH+: consequently,
holds for ironACTHUNTGRNEUNG(III) complexes: a comparable intermediate
can be formed according to Scheme 7, and the influence of
the oxidation state seems not to be decisive (see above). It
is therefore plausible that a complex [LFeACTHNUTRGNEUNG
(dbc)]+, as gener-
ated under the conditions outlined in entry 4, can show
mainly extradiol reactivity, and thus compares well, for in-
stance, with [TpiPr Fe
C
2
ering that Tp is also a facially coordinating tripodal ligand)
or with [(BnBPA)Fe
AHCTUNGTRENNUNG
pyridylmethyl)amine) dissolved in dichloromethane.[19b]
However, iron complexes with tetradentate ligands cannot
react analogously, since only two coordination sites are
available. This situation resembles that at the intradiol diox-
ygenase active site, where the His2Tyr2 motif provides four
protein ligands for ironACTHNUGRTENUNG(III), provided that the tyrosyl unit
coordinates after O2 binding, as shown in Scheme 4.[16] The
iron centre is then only capable of binding substrates and re-
action intermediates through two coordination sites, that is
Chem. Eur. J. 2009, 15, 5567 – 5576
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
5573

C. Limberg et al.
synthesize derivatives of L in
which the differences between
the steric requirements of the
individual donor functions are
more pronounced so that ste-
reochemical constraints are im-
posed more rigorously (for in-
stance, by introducing a tert-
butyl residue at the pyridyl
ring).
Experimental Section
General: Apart from the ligand syn-
thesis, all manipulations were carried
out in a glove box, or else by means
of Schlenk-type techniques involving
the use of a dry argon atmosphere.
The 1H and 13C NMR spectra were
recorded on a Bruker AV 400 NMR
spectrometer (1H 400.13 MHz; 13C
100.63 MHz) with CDCl3 as solvent
at 208C. The 1H and 13C NMR spectra
were calibrated against the residual
proton and natural abundance 13C
resonances of the deuterated solvent
CDCl3 (dH 7.26 ppm, dC 77.00 ppm).
Microanalyses were performed on a
Leco CHNS-932 elemental analyser
or on a HEKAtech EURO EA. ESI
mass spectra were recorded on an
Agilent Technologies 6210 time-of-
flight LC/MS. Infrared (IR) spectra
were recorded with a Digilab Excali-
bur FTS 4000 FTIR spectrometer
from samples prepared in KBr pellets.
Scheme 8. Proposed reaction mechanisms of catechol cleavage for five- and six-coordinate complexes.
to say, the strict pseudo-axial peroxide conformation is more
Solvents were dried using a Braun solvent purification system. (1-Methyl-
imidazol-2-yl)(pyridin-2-yl)methanone (I),[9] 3(5)-tert-butyl-pyrazole,[10] 1-
hydroxymethyl-3(5)-tert-butylpyrazole,[11] 3,5-di-tert-butyl-1-oxacyclohep-
ta-3,5-diene-2,7-dione (A),[33] and 4,6-di-tert-butyl-2-pyrone (B)[34] were
synthesized as described previously.
difficult to achieve and so acyl migration is favoured as op-
posed to alkenyl migration (Scheme 8).[16] Therefore, iron
complexes with tetradentate ligands also represent selective
catalysts for intradiol cleavage, and consistently the same is
true for [LFe
(dbc)Cl] (entry 1), in which, besides the triden-
CCDC-711192, 711193, 711194, 711195, 711196, and 718663 contain the
supplementary crystallographic data for this paper. These data can be ob-
tained free of charge from The Cambridge Crystallographic Data Centre
via www.ccdc.cam.ac.uk/data_request/cif.
tate ligand L, a chlorido ligand coordinates strongly to the
iron centre.
2-[1-Hydroxy(1-methyl-1H-imidazol-2-yl)(1-methyl-5-tert-butylpyrazol-3-
yl)methyl]pyridine (II): A 2.5m solution of nBuLi in hexane (17.0 mL,
0.042 mol, 2 equiv) was slowly added to a solution of 1-hydroxymethyl-
3(5)-tert-butylpyrazole (3.24 g, 0.021 mol) in THF at ꢀ788C. Following
the addition, the reaction mixture was stirred for 1 h at ꢀ788C, and then
for a further 1 h after warming to RT. After cooling to ꢀ608C, a solution
of (1-methylimidazol-2-yl)(pyridin-2-yl)methanone (I) (4.00 g, 0.021 mol,
1 equiv) in THF was added. The colour of the solution quickly turned
brown during this addition. After slowly warming to RT while stirring
overnight, the mixture was neutralized with dilute aqueous hydrochloric
acid and the phases were allowed to separate. The aqueous phase was ex-
tracted with dichloromethane (3ꢄ30 mL), and the combined organic
phases were concentrated to dryness. The crude product thus obtained
was dissolved in dichloromethane (10 mL) and the solution was overlay-
ered with pentane (30 mL). Within 1 d, II had precipitated in the form of
a white solid (3.30 g, 0.011 mol, 52%). 1H NMR (CDCl3, 258C): d=1.27
Conclusion
A novel ligand, L, has been synthesized that shows interest-
ing properties for application in bioinspired or biomimetic
chemistry, as exemplified here for the case of catechol diox-
ygenase mimicking. The results obtained for the system
LFe3+/DBCH2/O2 under various conditions can be explained
on the basis of the current concepts of understanding con-
cerning the behaviour of native IDOs and EDOs, as well as
the reactions of corresponding model complexes, and they
show the general potential of L to model a histidine-rich
facial triad in enzymes. Future research will be aimed at the
synthesis of enantiomerically pure complexes of ligand L
and their investigation with respect to enantioselective oxi-
dations of prochiral substrates. To this end, we intend to
(s, 9H; CACTHNUGTRNE(UNG CH3)3), 3.25 (s, 3H; Im N-CH3), 6.02 (s, 1H; Pz CH-4), 6.82 (d,
1H, J=1.20 Hz; Im CH-4/5), 6.90 (d, 1H, J=1.20 Hz; Im CH-4/5), 7.26
(m, 1H; Py CH), 7.68 (m, 1H; Py CH), 7.71 (m, 1H; Py CH), 8.54 ppm
(m, 1H; Py CH); 13C NMR (CDCl3, 258C): d=30.32 (C
ACTHUNGTRNEUNG(CH3)3), 31.42 (C-
5574
www.chemeurj.org
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 5567 – 5576

Biomimetic Oxidation
FULL PAPER
A
volume of 10 mL, which led to the deposition of a yellow-orange precipi-
tate of [LFeCl3], 3. The product was collected by filtration, washed with
THF (2ꢄ5 mL), and dried to afford pure 3 in the form of a yellow-
orange solid (740 mg, 1.44 mmol, 68%). IR (KBr): n˜ =3159 (m), 3138
(m), 3123 (m), 3100 (m), 3078 (m), 3016 (w), 2975 (m), 2971 (m), 2961
(m), 2937 (m), 2907 (m), 2870 (m), 2853 (m), 1599 (s), 1527 (m), 1490
(m), 1467 (m), 1440 (m), 1366 (m), 1274 (m), 1243 (m), 1213 (m), 1160
(w), 1131 (s), 1101 (m), 1093 (m), 1063 (m), 1051 (m), 1014 (w), 985 (m),
889 (m), 849 (m), 775 (m), 762 cmꢀ1 (m); UV/Vis (acetonitrile) lmax1 (e)=
313 nm (4310mꢀ1 cmꢀ1), lmax2 (e)=356 nm (3820mꢀ1 cmꢀ1); ESI-MS (pos.
mode, MeCN): m/z: calcd for [MꢀCl]+: 465.0786; found: 465.0777; ele-
mental analysis calcd (%) for 3·0.5CH2Cl2 (C19.5H26Cl4FeN5O): C 43.04,
H 4.82, N 12.87, Cl 26.06; found C 43.47, H 4.90, N 12.79, Cl 25.05.
(Im CH-4/5), 126.32 (Py CH), 137.02 (Py CH), 146.61 (Py CH),
159.23 ppm (Py CH); MS: m/z: 311 [M+], 296, 278, 254, 233, 205, 188,
177, 151, 122, 106, 83; elemental analysis calcd (%) for C17H21N5O: C
65.57, H 6.80, N 22.49; found: C 65.18, H 6.83, N 22.48.
2-[1-Methoxy(1-methylimidazol-2-yl)(1-methyl-5-tert-butylpyrazol-3-yl)-
methyl]pyridine (L): NaH (850 mg, 35.4 mmol, 5.2 equiv) was added por-
tionwise to a solution of the alcohol II (2.11 g, 6.77 mmol) in THF. The
reaction mixture was stirred under argon for 7 d at RT and then heated
to reflux for 5 h. After cooling to RT, iodomethane (1.5 mL, 24.1 mmol,
3.6 equiv) was added, and after stirring for a further 5 h all volatiles were
removed under vacuum. Excess NaH was decomposed by treatment with
water, and the resulting mixture was extracted with dichloromethane (3ꢄ
20 mL). Removal of the solvent from the combined extracts left the
crude product as a yellow-brown solid, which could be purified by wash-
ing with THF to afford L in the form of a white powder (1.30 g,
Oxidative cleavage of 3,5-di-tert-butylcatechol (DBCH2): In a typical re-
action, 3 (45 mg, 0.09 mmol), DBCH2 (20 mg, 0.09 mmol, 1 equiv), NEt3
(25 mL, 0.18 mmol, 2 equiv), and sodium tetraphenylborate (31 mg,
0.09 mmol, 1 equiv) were dissolved in acetonitrile (30 mL) under an inert
argon atmosphere in a glove box, and the reaction mixture was stirred
for 5 min. Thereafter, excess O2 was introduced and the mixture was
stirred for a further 6 h. All volatiles were then removed under vacuum,
and the residue was redissolved in CHCl3. After filtration of this solution
through silica gel, all volatiles were removed once more. One equivalent
of naphthalene was added as a standard and the residue was analyzed by
1H NMR spectroscopy, GC-MS, and GC-FID.[35] Data for 3,5-di-tert-bu-
tylquinone (Q): 1H NMR (CDCl3, 258C): d=1.23 (s, 9H), 1.28 (s, 9H),
6.22 (d, 1H), 6.93 ppm (d, 1H); data for 3,5-di-tert-butyl-1-oxacyclohep-
3.83 mmol, 57%). 1H NMR (CDCl3, 258C): d=1.32 (s, 9H; C
ACTHNUGTRNEUGN(CH3)3),
3.28 (s, 3H; Im N-CH3), 3.31 (s, 3H; O-CH3), 3.90 (s, 3H; Pz N-CH3),
6.28 (s, 1H; Pz CH-4), 6.85 (d, 1H, J=1.19 Hz; Im CH-4/5), 7.05 (d, 1H,
J=1.16 Hz; Im CH-4/5), 7.12 (m, 1H; Py CH), 7.66 (m, 1H; Py CH),
7.76 (m, 1H; Py CH), 8.53 ppm (m, 1H; Py CH); 13C NMR (CDCl3,
258C): d=29.55 (CACHTUNGTRENNUNG(CH3)3), 31.06 (CACHTUGNTREN(NUNG CH3)3), 34.02 (Im N-CH3), 39.45 (Pz
N-CH3), 52.96 (O-CH3), 81.57 (Cq-OCH3), 104.62 (Pz CH-4), 121.85 (Py
CH), 122.90 (Py CH), 122.81 (Im CH 4/5), 126.48 (Im CH 4/5), 136.12
(Py CH), 146.76 (Im C-2), 148.34 (Py CH), 149.50 (Pz C-3), 150.96 (Pz C-
5), 161.69 ppm (Py C-2); IR (KBr): n˜ =3050 (m), 2996 (m), 2974 (s), 2961
(s), 2950 (s), 2932 (m), 2927 (m), 2923 (m), 2909 (m), 2890 (m), 2887 (m),
2871 (m), 2821 (m), 1586 (m), 1527 (m), 1488 (m), 1468 (s), 1437 (m),
1365 (m), 1281 (m), 1248 (m), 1214 (m), 1135 (w), 1094 (m), 1060 (s),
1022 (w), 994 (w), 957 (m), 887 (m), 811 (m), 801 (m), 780 (s), 758 (s),
738 (m), 725 (m), 699 (w), 666 (w), 619 cmꢀ1 (w); ESI-MS (pos. mode,
MeCN): m/z: calcd for MH+: 340.2137; found: 340.2127; elemental analy-
sis calcd (%) for C19H25N5O: C 67.23, H 7.42, N 20.63; found C 67.26, H
7.62, N 20.94.
1
ta-3,5-diene-2,7-dione (A): H NMR (CDCl3, 258C): d=1.17 (s, 9H), 1.29
(s, 9H), 6.15 (d, 1H), 6.45 ppm (d, 1H); CI-MS (methanol): m/z: 237.1
[M+H+]; data for 4,6-di-tert-butyl-2-pyrone (B): 1H NMR (CDCl3,
258C): d=1.23 (s, 9H), 1.29 (s, 9H), 6.05 ppm (m, 2H); EI-MS: m/z=
208.1 [M+]; CI-MS (methanol): m/z: 209.2 [M+H+]; data for 3,5-di-tert-
butyl-2-pyrone (C): 1H NMR (CDCl3, 258C, ppm): d=1.22 (s, 9H), 1.37
(s, 9H), 7.21 (d, 1H), 7.23 ppm (d, 1H); EI-MS: m/z: 208.1 [M+]; CI-MS
(methanol): m/z: 209.2 [M+H+].
ACHTUNGTRENNUNG[LFeCl2] (1): L (800 mg, 2.36 mmol, 1 equiv) was added to a suspension
of FeCl2 (299 mg, 2.36 mmol) in THF (20 mL). The reaction mixture was
stirred overnight and then the solution obtained was concentrated to a
volume of 10 mL, which led to the deposition of a white precipitate of
[LFeCl2], 1. The product was collected by filtration, washed with THF
(2ꢄ5 mL), and dried to afford pure 1 in the form of a white solid
(670 mg, 1.44 mmol, 61%). Crystals suitable for single-crystal X-ray dif-
fraction studies could be obtained by evaporating the solvent from satu-
rated solutions in THF or acetonitrile. According to the XRD results,
these crystals contained one solvent molecule per two molecules of 1.
The crystals grown in THF were of better quality, and so Figure 1 relates
to an investigation of these crystals. IR (KBr): n˜ =3123 (m), 2974 (s),
2937 (m), 2913 (m), 2911 (m), 2908 (m), 2898 (m), 2866 (m), 2852 (w),
1586 (m), 1533 (m), 1496 (m), 1464 (m), 1431 (m), 1424 (m), 1377 (w),
1370 (m), 1246 (m), 1154 (m), 1101 (m), 1081 (s), 1071 (s), 1047 (w), 980
(m), 906 (m), 811 (m), 807 (m), 783 (s), 775 (m), 768 (m), 765 (m), 753
(m), 722 (w), 714 (m), 704 cmꢀ1 (w); ESI-MS (pos. mode, MeCN): m/z:
calcd for [MꢀCl]+: 430.1097; found: 430.1090; elemental analysis calcd
Acknowledgements
We are grateful to the Fonds der Chemischen Industrie, the Bundesminis-
terium fꢀr Bildung, Wissenschaft, Forschung und Technologie, and the
Dr. Otto Rçhm Gedꢁchtnisstiftung for financial support, as well as to
Bayer GmbH, BASF AG, and Sasol GmbH for the supply of chemicals.
We would also like to thank E. Hoppe, C. Knispel, P. Neubauer, and B.
Ziemer for performing the crystal structure analyses, Prof. P. Wessig for
carrying out analytical HPLC investigations, and C. Jankowski for prepar-
ing the starting materials. Finally, we acknowledge inspiring discussions
with Professor Peter Comba and within the Cluster of Excellence “Unify-
ing Concepts in Catalysis” funded by the DFG.
[1] H. Decker, T. Schweikardt, F. Tuczek, Angew. Chem. 2006, 118,
4658–4664; Angew. Chem. Int. Ed. 2006, 45, 4546–4550.
[2] W. N. Lipscomb, N. Strꢁter, Chem. Rev. 1996, 96, 2375–2433.
[3] A. L. Feig, S. J. Lippard, Chem. Rev. 1994, 94, 759–805.
[4] S. J. Lange, L. Que, Jr., Curr. Op. Chem. Biol. 1998, 2, 159–172; A.
Beck, B. Weibert, N. Burzlaff, Eur. J. Inorg. Chem. 2001, 521–527;
M. M. Abu-Omar, A. Loaiza, N. Hontzeas, Chem. Rev. 2005, 105,
2227–2252; D. Oldenburg, C. Y. Ke, A. Tipton, A. A. Shteinman, L.
Que, Jr., Angew. Chem. 2006, 118, 8143–8146; Angew. Chem. Int.
Ed. 2006, 45, 7975–7978.
[5] S. Trofimenko, Chem. Rev. 1993, 93, 943–980.
[6] L. Zhou, D. Powell, K. M. Nicholas, Inorg. Chem. 2007, 46, 2316–
2321.
[7] a) G. Anderegg, F. Wenk, Helv. Chim. Acta 1967, 50, 2330–2332;
b) A. Diebold, K. S. Hagen, Inorg. Chem. 1998, 37, 215–223; S. Yan,
D. D. Cox, L. L. Pearce, C. Juarez-Garcia, L. Que, Jr., J. H. Zhang,
C. J. OꢅConnor, Inorg. Chem. 1989, 28, 2507–2509.
(%) for crystals of
1 grown from CH3CN, that is, 1·0.5CH3CN
(C20H26.5Cl2FeN5.5O): C 49.35, H 5.49, N 15.83, Cl 14.57; found C 49.37, H
5.46, N 15.61, Cl 14.62.
ACHTUNGTRENNUNG[LFeBr2] (2): L (300 mg, 0.88 mmol, 1 equiv) was added to a solution of
FeBr2 (191 mg, 0.88 mmol) in THF (20 mL). The reaction mixture was
stirred overnight and then the solution obtained was concentrated to a
volume of 10 mL, which led to the deposition of a white precipitate of
[LFeBr2], 2. The product was collected by filtration, washed with THF
(2ꢄ5 mL), and dried to afford pure 2 in the form of a white solid
(230 mg, 0.41 mmol, 47%); ESI-MS (pos. mode, MeCN): m/z: calcd for
[MꢀBr]+: 474.06383; found: 474.0581; elemental analysis calcd (%) for
2·0.5THF (C21H29Br2FeN5O1.5): C 42.67, H 4.94, N 11.85, Br 27.03; found
C 42.80, H 5.03, N 11.78, Br 24.87.
ACHTUNGTRENNUNG[LFeCl3] (3): L (740 mg, 2.18 mmol, 1 equiv) was added to a solution of
FeCl3 (354 mg, 2.18 mmol) in THF (20 mL). The reaction mixture was
stirred overnight and then the solution obtained was concentrated to a
Chem. Eur. J. 2009, 15, 5567 – 5576
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
5575

C. Limberg et al.
[8] I. G. Denisov, T. M. Makris, S. G. Sligar, I. Schlichting, Chem. Rev.
2005, 105, 2253–2277; P. R. Ortiz de Montellano, in Cytochrome
P450—Structure, Mechanism, and Biochemistry (Ed.: P. R. Ortiz de
Montellano), Plenum Press, New York, 1995, pp. 245–303.
[9] M. Costas, M. P. Mehn, M. P. Jensen, L. Que, Jr., Chem. Rev. 2004,
104, 939–986.
[10] A. J. Canty, E. E. George, C. V. Lee, Aust. J. Chem. 1983, 36, 415–
418.
[11] S. Trofimenko, J. C. Calabrese, J. S. Thompson, Inorg. Chem. 1987,
26, 1507–1514.
[12] R. Hꢀttel, P. Jochum, Chem. Ber. 1952, 85, 820–826.
[13] M. Wagner, C. Limberg, Eur. J. Inorg. Chem. 2008, 3970–3976;
C. D. Burbridge, D. M. L. Goodgame, J. Chem. Soc. A 1968, 1074–
1079; S. Hikichi, T. Ogihara, K. Fujisawa, N. Kitajima, M. Akita, Y.
Moro-oka, Inorg. Chem. 1997, 36, 4539–4547; J. L. Kisko, T. Hascall,
G. Parkin, J. Am. Chem. Soc. 1998, 120, 10561–10562.
[14] a) R. G. McDonald, M. A. Hitchman, Inorg. Chem. 1990, 29, 3074;
b) T. Astley, A. J. Canty, M. A. Hitchman, G. L. Rowbottom, B. W.
Skelton, A. H. White, J. Chem. Soc. Dalton Trans. 1991, 1981.
[15] R. M. Acheson, An Introduction to the Chemistry of Heterocyclic
Compounds, Wiley, New York, 1976, p. 335.
Koch, H. J. Krꢀger, Angew. Chem. 1995, 107, 2928–2931; Angew.
Chem. Int. Ed. Engl. 1996, 34, 2671–2674; g) M. Velusamy, R.
Mayilmurugan, M. Palaniandavar, Inorg. Chem. 2004, 43, 6284–
6293.
[21] D. D. Cox, S. J. Benkovic, L. M. Bloom, F. C. Bradley, M. J. Nelson,
L. Que, Jr., D. E. Wallick, J. Am. Chem. Soc. 1988, 110, 2026–2032.
[22] T. Funabiki, Y. Nagai, H. Kojima, T. Tanaka, S. Yoshida, H.
Masuda, Inorg. Chim. Acta 1998, 275–276, 222–229.
[23] G. Ackermann, D. Hesse, Z. Anorg. Chem. 1969, 368, 25–30.
[24] E. Mentasti, E. Pelizzetti, G. Saini, J. Chem. Soc. Dalton Trans.
1973, 2609.
[25] R. Mayilmurugan, H. Stoeckli-Evans, M. Palaniandavar, Inorg.
Chem. 2008, 47, 6645–6658.
[26] S. Mendel, A. Arndt, T. D. H. Bugg, Biochemistry 2004, 43, 13390–
13396.
[27] S. L. Groce, J. D. Lipscomb, Biochemistry 2005, 44, 7175–7188.
[28] M. W. Vetting, L. P. Wackett, L. Que, Jr., J. D. Lipscomb, D. H. Oh-
lendorf, J. Bacteriol. 2004, 186, 1945–1958.
[29] J. S. Olson, A. J. Mathews, R. J. Rohls, B. A. Springer, K. D. Ege-
berg, S. G. Sligar, J. Tame, J. P. Renaud, K. Nagai, Nature 1988, 336,
265–266.
[16] T. D. H. Bugg, G. Lin, Chem. Commun. 2001, 941–952.
[17] Y. M. Chiou, L. Que, Jr., Inorg. Chem. 1995, 34, 3577–3578; D. H.
Jo, Y. M. Chiou, L. Que, Jr., Inorg. Chem. 2001, 40, 3181–3190.
[18] a) P. C. A. Bruijnincx, M. Lutz, A. L. Spek, W. R. Hagen, G. von Ko-
ten, R. J. M. Klein Gebbink, J. Am. Chem. Soc. 2007, 129, 2275–
2286; b) T. Ogihara, S. Hikichi, M. Akita, Y. Moro-oka, Inorg.
Chem. 1998, 37, 2614–2615.
[30] The only synthetic complex that exclusively affords extradiol prod-
ucts is [Fe(Cl)ACHTNUTGRENN(GU dtbc)ACHTUNTGERN(NUGN Me3-tacn)], provided that 1 equiv of AgBF4 is
added to generate a vacant coordination site.[19b] This system seems
to be quite exceptional, as in most cases both product types are
found.
[31] M. Fujiwara, L. A. Golovleva, Y. Saeki, M. Nozaki, O. Hayashi, J.
Biol. Chem. 1975, 250, 4848–4855.
[19] a) M. Ito, L. Que, Jr., Angew. Chem. 1997, 109, 1401–1403; Angew.
Chem. Int. Ed. Engl. 1997, 36, 1342–1344; b) D. H. Jo, L. Que, Jr.,
Angew. Chem. 2000, 112, 4454–4457; Angew. Chem. Int. Ed. 2000,
39, 4284–4287; c) P. C. A. Bruijnincx, M. Lutz, A. L. Spek, W. R.
Hagen, G. von Koten, R. J. M. Klein Gebbink, Inorg. Chem. 2007,
46, 8391–8402; d) K. Visvaganesan, R. Mayilmurugan, E. Suresh,
M. Palaniandavar, Inorg. Chem. 2007, 46, 10294–10306.
[20] a) L. Que, Jr., R. C. Kolanczyk, L. S. White, J. Am. Chem. Soc. 1987,
109, 5373–5380; b) D. D. Cox, L. Que, Jr., J. Am. Chem. Soc. 1988,
110, 8085–8092; c) H. G. Jang, D. D. Cox, L. Que, Jr., J. Am. Chem.
Soc. 1991, 113, 9200–9204; d) M. Duda, M. Pascaly, B. Krebs, Chem.
Commun. 1997, 835–836; e) P. Mialane, L. Tchertanov, F. Banse, J.
Sainton, J. J. Girerd, Inorg. Chem. 2000, 39, 2440–2444; f) W. O.
[32] J. Schlosrich, K. L. Eley, P. J. Crowley, T. D. H. Bugg, ChemBioChem
2006, 7, 1899–1908.
[33] T. R. Demmin, M. M. Rogic, J. Org. Chem. 1980, 45, 1153–1156.
[34] Y. Tatsuno, M. Tatsuda, S. Otsuka, J. Chem. Soc. Chem. Commun.
1982, 1100–1101.
[35] For the GC-MS/FID analysis, it is important that all NaBPh4 is re-
moved during the work-up procedure, as otherwise a reaction with
any Q present in the sample occurs within the injector unit to give
the boric ester (DBC)BPh. This would lead to a falsely suppressed
yield of Q.
Received: December 10, 2008
Revised: February 19, 2009
Published online: April 9, 2009
5576
www.chemeurj.org
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 5567 – 5576