Reactions of tetraazamacrocyclic FEIII complexes with hydrogen peroxide - Putative catalase mimics?

Article abstract of DOI:10.1002/1099-0690(200108)2001:16<3119::AID-EJOC3119>3.0.CO;2-T

Iron(III) complexes of macrocyclic tetraaza ligands [Fe(III)-1-Fe(III)-3] were investigated for their putative catalase-like properties under physiological conditions: i.e., in aqueous solution at pH = 7.0-7.4 and at micromolar concentrations of the catal

Full text of DOI:10.1002/1099-0690(200108)2001:16<3119::AID-EJOC3119>3.0.CO;2-T

FULL PAPER
Reactions of Tetraazamacrocyclic FeIII Complexes with Hydrogen Peroxide ؊
Putative Catalase Mimics?
Sabrina Autzen,[a] Hans-Gert Korth,[a] Herbert de Groot,*[b] and Reiner Sustmann*[a]
Keywords: Catalase / Enzyme models / Peroxides / Iron / Macrocyclic ligands
Iron(III) complexes of macrocyclic tetraaza ligands [Fe(III)-
1−Fe(III)-3] were investigated for their putative catalase-like
properties under physiological conditions: i.e., in aqueous so-
lution at pH = 7.0−7.4 and at micromolar concentrations of
the catalyst and hydrogen peroxide. Complex Fe(III)-1, ori-
ginally studied by Busch et al. as a catalase model, at pH =
4.6, only degrades hydrogen peroxide to oxygen as a minor
reaction at this pH (Ͻ 1% O2 yield). Experiments with the
analogous complex Fe(III)-2 at pH = 7.2 found that no oxy-
gen was formed under physiological conditions, although hy-
drogen peroxide was decomposed to an extent of more than
50%. Complex Fe(III)-3 produced oxygen on reaction with
H2O2, but only in stoichiometric amounts. Thus, the decom-
position of hydrogen peroxide by these iron(III) complexes
cannot reasonably be described as a catalase-like activity.
Introduction
simultaneously with the ferryl species. The catalytic cycle is
concluded by reaction between the ferryl ironϪporphyrin
radical cation (Compound I in enzymologists’ terminology)
and a second molecule of hydrogen peroxide. Oxygen is lib-
erated and iron(III)Ϫporphyrin is regenerated. The mech-
anism through which manganese-containing enzymes des-
troy hydrogen peroxide cannot be the same, since these en-
zymes contain two manganese ions in their active centers.
Attempts to synthesize both iron and manganese com-
The preparation of synthetic enzymes is a challenge, be-
cause low molecular mass model compounds are expected
to catalyze the same reaction as the enzyme, hopefully by
the same mechanism and with high activity and efficiency.
Catalases belong to the most active enzymes and in nature
serve to protect living cells from the toxicity of hydrogen
peroxide,[1,2] a ubiquitous metabolite produced by various
oxidases and superoxide dismutases. Hydrogen peroxide plexes as catalase mimics Ϫ that is, compounds that decom-
seems to perform regulatory functions at low concentra- pose hydrogen peroxide into water and oxygen Ϫ have been
tions, but causes cell damage at higher levels, especially if made. So far, more success seems to have been achieved
hydroxyl radicals are generated through interaction with with organometallic manganese compounds, in particular
low-valent metal ions such as iron(II).[3] To prevent this MnϪsalen complexes.[6] It has long been known that hydro-
damage, cells make use of catalases or peroxidases to neut- gen peroxide is decomposed catalytically in the presence of
ralize hydrogen peroxide, in a process in which two hydro- iron complexes. The chemistry of iron(II)/iron(III) with hy-
gen peroxide molecules are transformed into two molecules drogen peroxide in aqueous solution is rather complex and
of water and one molecule of oxygen. Eucaryotic and pro- the mechanistic details are still debated.[7Ϫ11] Less common
caryotic cells mainly use enzymes containing FeIII
Ϫ
are studies in which FeIII complexes are viewed as catalase
porphyrin systems for this purpose. A few manganese com- mimics able to decompose hydrogen peroxide efficiently,
plexes are found as active components[3] in catalases from and perhaps by a mechanism similar to that of the
procaryotic cells.
enzyme.[12Ϫ14] In 1979, Melnyk et al.[15] introduced a non-
Several catalases containing iron(III)Ϫporphyrin units heme tetraazamacrocyclic FeIII complex [Fe(III)-1], for
have been isolated and structurally analyzed by X-ray crys- which oxygen evolution was observed on interacting with
tallography,[1] and the mechanism through which hydrogen hydrogen peroxide under acidic conditions. An account on
peroxide is decomposed has been established.[4,5] The prim- previous studies of decomposition of hydrogen peroxide by
ary step is the interaction of one molecule of hydrogen per- iron(III) compounds is given in that publication. The same
oxide with the iron(III) center of the porphyrin unit to give group later[16] proposed that this catalytic activity proceeds
a ferryl iron complex (FeIVϭO). As hydrogen peroxide is a through a high oxidation state Fe intermediate rather than
two-electron oxidant, a porphyrin radical cation is formed through involvement of free hydroxyl radicals. The complex
therefore had catalase-like and peroxidase-like activity at-
[a]
Institut für Organische Chemie, Universität Essen,
Universitätsstrasse 5, 45117 Essen, Germany
Fax: (internat.) ϩ 49-(0)201/183-4259
tributed to it. Recently, Busch’s group also described a
newly designed substituted triazacyclononaneϪFeII com-
plex which decomposes hydrogen peroxide by a complex,
multistep mechanism.[16,17]
Our goal is the development of low molecular mass FeIII
complexes that are soluble in water and decompose hydro-
E-mail: sustmann@oc1.orgchem.uni-essen.de
Institut für Physiologische Chemie, Universitätsklinikum Essen,
Hufelandstrasse 55, 45122 Essen, Germany
Fax: (internat.) ϩ 49-(0)201/723-5943
[b]
E-mail: h.de.groot@uni-essen.de
Eur. J. Org. Chem. 2001, 3119Ϫ3125
WILEY-VCH Verlag GmbH, 69451 Weinheim, 2001
1434Ϫ193X/01/0816Ϫ3119 $ 17.50ϩ.50/0
3119

S. Autzen, H.-G. Korth, H. de Groot, R. Sustmann
FULL PAPER
Results and Discussion
Synthesis of Ligands and FeIII Complexes
Ligands 1 and 2 were prepared according to literature
procedures.[16,18] The preparation of ligand 3 was devised
similarly, starting from pyridine-2,6-dicarbaldehyde; this
was treated with the appropriate diamine in a copper tem-
plate reaction to give the macrocycle. Diimine reduction by
sodium borohydride and removal of copper(II) as sulfide
provided 3 in 81% overall yield. In native catalase, a tyros-
ine residue occupies the distal site of the
iron(III)Ϫporphyrin unit. In order to simulate this situ-
ation, ligands bearing 2-methoxy or 2-hydroxybenzyl
groups at the nitrogen atom opposite to the pyridine nitro-
gen atom (4a, 4b) were envisaged. The synthesis of these
ligands started from 2-methoxybenzylamine, which was ad-
ded to 2 equiv. of acrylonitrile in a Michael-type reaction.
Reduction of the dinitrile to the diamine provided the
building block, which could be subjected to macrocycliz-
ation as in the case of ligands 1Ϫ3.
An iron(III) complex had so far been known only for 1
[Fe(III)-1]. This complex was synthesized according to the
procedure of Melnyk et al.[15] Compounds Fe(III)-2 and
Fe(III)-3 were prepared similarly, but with addition of tetra-
ethylammonium tetrafluoroborate instead of tetrafluoro-
boric acid after air oxidation. Unfortunately, attempts to
prepare iron(III) complexes of ligands 4 were unsuccessful,
although a number of different experimental conditions
were tested.
Reactions of Fe(III)-1؊Fe(III)-3 with Hydrogen Peroxide
Rates of reaction of complex Fe(III)-1 with hydrogen per-
oxide had been studied thoroughly as functions of the ini-
tial concentration of the catalyst (72 µϪ1.5 m), hydro-
gen peroxide concentration (68Ϫ360 m), and pH
(3.94Ϫ5.10).[15] Rate constants were obtained from initial
rates of oxygen production, determining the formation of
oxygen from the pressure increase in a closed reaction sys-
tem. Unfortunately, Melnyk et al.[15] did not mention the
total yield of oxygen formed under their conditions. The
only information given is that oxygen production satisfies
the stoichiometry implied by the catalase mechanism. We
Scheme 1
gen peroxide to oxygen and water in high yield under were able to reproduce these results by performing experi-
physiological conditions (i.e. at pH ϭ 7.0Ϫ7.4) and at mic- ments with [H2O2] ϭ 260 m and [Fe(III)-1] ϭ 1 m at
romolar concentrations of both catalyst and hydrogen per- pH ϭ 4.6, using a Toepler pump to determine the volume
oxide. Such compounds should be attractive for biological of released oxygen. A turn-over number (TON) of [O2]/
applications. Porphyrin systems, on which many model [Fe(III)-1] ϭ 50 was determined.
studies have been carried out, were discounted, mainly be-
When we reduced the concentration of Fe(III)-1 to values
cause of their low solubility in aqueous solution, making closer to iron levels typically found in physiological systems,
them inappropriate for biological investigations. As a start- however, the production of oxygen, monitored electrochem-
ing point for our work we decided to reinvestigate the cata- ically by a Clark electrode, was insignificant relative to the
lase-like activity of the FeIII complex of 1 [Fe(III)-1] under amount of oxygen calculated for 100% conversion of hydro-
physiological conditions and to progress from this com- gen peroxide. Figure 1 (a) shows the yield of oxygen as a
pound to related structures that might be suitable as cata- function of [Fe(III)-1] at [H2O2] ϭ 260 m at pH ϭ 4.6. It
lase mimics. Here we report on our investigations on the can be seen that no significant oxygen production was de-
non-annulene-type tetraazamacrocyclic FeIII complexes tected at [Fe(III)-1] Ͻ 10 µ. An increase of the amount of
Fe(III)-1ϪFe(III)-3.
catalyst resulted in an increase in oxygen formation, but in
3120
Eur. J. Org. Chem. 2001, 3119Ϫ3125

Reactions of Tetraazamacrocyclic FeIII Complexes with Hydrogen Peroxide
FULL PAPER
comparison with the total yield that would be possible on conversion of hydrogen peroxide, assuming the catalase
the basis of the catalase stoichiometry (100% ϭ 130 m stoichiometry. At 5 m H2O2 and at the same catalyst con-
O2), the formation of oxygen was less than 1%.
centration, the yield of oxygen is slightly increased, to about
5.4%. These experiments clearly demonstrate that oxygen
formation, which depends both on catalyst and hydrogen
peroxide concentration, is only a minor process.
Figure 2. Oxygen yield as a function of [Fe(III)-2]; [H2O2] ϭ 40 m,
closed circles; [H2O2] ϭ 5 m; open circles; T ϭ 25 °C, pH ϭ 4.6
Complex Fe(III)-2 behaves virtually identically to Fe(III)-
1, as is evident from Figure 3, in which oxygen formation
was measured as a function of H2O2 concentration in the
presence of 22 µ Fe(III)-2, comparing this with the data
for the same reaction of Fe(III)-1. The measured yields of
oxygen (average from slope: 0.42%) for the reaction between
Fe(III)-1 and hydrogen peroxide (closed circles) are virtually
identical to the values (average from slope: 0.36%) found
for the same reaction with Fe(III)-2 (open circles).
Figure 1. Oxygen yield from the reaction between Fe(III)-1 and
H2O2 at T ϭ 25 °C and pH ϭ 4.6 as: a) a function of [Fe(III)-1]
([H2O2] ϭ 260 m); b) a function of [H2O2] ([Fe-1] ϭ 22 µ)
A similar conclusion follows from Figure 1 (b), in which
oxygen production is determined as a function of the con-
centration of H2O2 at an Fe(III)-1 concentration of 22 µ.
From the slope of the regression line, it is possible to calcu-
late an average relative yield of O2 of 0.4% for this range of
hydrogen peroxide concentration. Thus, although oxygen is
formed by the reaction between Fe(III)-1 and hydrogen per-
oxide at acidic pH,[15,16] this reaction is insignificant relative
to the extent of hydrogen peroxide degradation.
Experiments similar to those with Fe(III)-1 were carried
out with complex Fe(III)-2, which differs from Fe(III)-1 in
the absence of the methyl groups in the macrocycle. Fig-
ure 2 displays the yield of oxygen as a function of the con-
centration of Fe(III)-2 from 10 to 100 µ, for two different
Figure 3. Comparison of the total yield of oxygen from the reac-
tions of Fe(III)-1 (closed circles) and Fe(III)-2 (open circles) with
H2O2 as a function of [H2O2]
More important than the formation of oxygen at pH ϭ
hydrogen peroxide concentrations (40 m, upper line; 4.6, however, is the question of whether oxygen is produced
5 m, lower line). At both concentrations, the amount of at a physiological pH of 7.2, a point not addressed in
oxygen increases linearly with the concentration of the cata- ref.[15,16] This property would make an enzyme mimic useful
lyst. At 40 m H2O2 and at 100 µ Fe(III)-2, the yield of for applications in cell biology studies. Experiments ad-
oxygen is only 1.2% of the total yield expected for 100% dressing this question were carried out with complex
Eur. J. Org. Chem. 2001, 3119Ϫ3125
3121

S. Autzen, H.-G. Korth, H. de Groot, R. Sustmann
1
FULL PAPER
Fe(III)-2. This complex, however, does not produce detect- Exploratory H NMR experiments in this direction were
able (Ն 0.1 µ) amounts of oxygen at pH ϭ 7.2 and performed. After termination of the reaction between
[Fe(III)-2] ϭ 22 µ. Figure 4 displays the amount of hydro- Fe(III)-2 and H2O2, several new signals that indicated an
gen peroxide decomposed in the presence of [Fe(III)-2] ϭ unspecific attack on the ligand were observed. The chemical
22 µ as a function of hydrogen peroxide concentration. shifts of the new signals point to oxygenation at the ali-
The straight line indicates theoretical 100% decomposition phatic and aromatic positions, as well as cleavage of the
of hydrogen peroxide. The experiment was carried out in ring. The formation of oxygenated products would account
such a way that, 150 s after addition of Fe(III)-2 to the for the loss of hydrogen peroxide without release of oxygen.
buffer solution of hydrogen peroxide, the unchanged hydro-
A change in reactivity towards H2O2 was presumed for
gen peroxide was determined by addition of native catalase Fe(III)-3, due to the presence of two aromatic residues, each
to the reaction mixture. The amount of hydrogen peroxide directly attached to an amino group of the macrocycle. Fig-
decomposed was then calculated from the amount of oxy- ure 5 demonstrates that oxygen was produced in the reac-
gen released by catalase. Control experiments revealed that tion between Fe(III)-3 and H2O2. At an H2O2 concentration
after 150 s no further decomposition of hydrogen peroxide of 500 µ, a linear increase of oxygen formation as a func-
took place. It is interesting that up to 50% of hydrogen per- tion of the concentration of Fe(III)-3 is observed. Produc-
oxide was decomposed, although no oxygen production tion of O2 (0Ϫ11%), however, was much smaller than ex-
could be detected under these conditions. Figure 4 indicates pected for a quantitative catalytic decomposition of H2O2
that the presence of superoxide dismutase (SOD) has no with respect to the catalase stoichiometry. Experiments with
effect on the amount of H2O2 decomposed, and nor is any added catalase proved that the released O2 fitted well with
oxygen formed under these conditions. These facts largely the amount of decomposed hydrogen peroxide, or in other
rule out the intermediacy of superoxide radical anion in the words, no further decomposition of H2O2 is initiated by
decomposition process. Thus, Fe(III)-2 is an active catalyst complex Fe(III)-3, unlike the cases of complexes Fe(III)-1
for hydrogen peroxide degradation at pH ϭ 7.2 but cannot and Fe(III)-2. Inspection of the data of Figure 5 (slope of
reasonably be considered to be a catalase mimic.
regression line ϭ 0.51) reveals that the reaction is stoichi-
ometric, with one molecule of oxygen being produced per
two molecules of Fe(III)-3.
Figure 4. Decomposition of hydrogen peroxide without formation
of oxygen in the presence of Fe(III)-2, as a function of [H2O2]
([Fe(III)-2] ϭ 22 µ; T ϭ 25 °C; pH ϭ 7.2), determined as the
difference between the 100% value and the amount of unchanged
H2O2 (which was measured as oxygen after addition of catalase);
broken line: decomposition in the presence of SOD
Figure 5. Oxygen yield from the reaction between H2O2 and
Fe(III)-3 as a function of [Fe(III)-3] ([H2O2] ϭ 500 µ; T ϭ 25 °C;
pH ϭ 7.2)
In order to provide preliminary evidence for the mechan-
ism of H2O2 decomposition, some exploratory experiments
were performed in the presence of 2,2Ј-azinobis(3-ethyl-1,2-
dihydrobenzthiazole-6-sulfonate) (ABTS),[19] which is fre-
Conclusion
Reinvestigation of the catalase activity of tetraazamacro-
quently used to identify hydroxyl radicals in enzymatic reac- cyclic FeIII complex Fe(III)-1 confirmed that Fe(III)-1 does
tions. It was concluded that free hydroxyl radicals are not produce oxygen at pH ϭ 4.6 as reported previously.[15]
released in the Fe(III)-2-promoted decomposition of H2O2. However, it turned out that this reaction is only a minor
Because of the high reactivity of hydroxyl radicals, however, side reaction (Ͻ 1%). Studies with the analogous complex
there exists the possibility that these species react directly Fe(III)-2 showed that no oxygen is formed at a physiological
with the ligand of Fe(III)-2 in a ‘‘site-specific’’ manner, thus pH of 7.2, despite the fact that hydrogen peroxide is decom-
being responsible for the destruction of the catalytically act- posed by an extent of more than 50% in the presence of
ive complex. Destruction of the catalytic complex by high- Fe(III)-2 under these conditions. Complex Fe(III)-3 does
valent oxoϪiron intermediates is also a possible scenario. produce oxygen on interaction with hydrogen peroxide, but
3122
Eur. J. Org. Chem. 2001, 3119Ϫ3125

Reactions of Tetraazamacrocyclic FeIII Complexes with Hydrogen Peroxide
FULL PAPER
temperature overnight. Copper(II) was removed by treating the
mixture with sodium sulfide nonahydrate (3.66 g, 15.2 mmol), fol-
lowed by stirring at room temperature for 1 h, and heating at 60
°C for 1 h. The solution was cooled, and the precipitate isolated
and dissolved in hot dichloromethane. Insoluble material was re-
moved by filtration. The dichloromethane solution was cooled to
10 °C to give light yellow needles of 3 (1.79 g, 5.43 mmol, 81%). Ϫ
M.p. 221 °C. Ϫ 1H NMR (500 MHz, [D6]DMSO): δ ϭ 2.24 (t,
3J ϭ 6.3 Hz, 1 H, NH), 3.74 (d, 3J ϭ 6.3 Hz, 4 H, 10-H, 12-H),
4.34 (d, 3J ϭ 4.4 Hz, 4 H, 2-H, 20-H), 6.57 (dd, 3J ϭ 7.3, 3J ϭ
7.3 Hz, 2 H, 7-H, 15-H), 6.60 (d, 3J ϭ 7.9 Hz, 2 H, 5-H, 17-H),
7.11 (m, 6 H, 6-H, 8-H, 14-H, 16-H, 2 ϫ NH), 7.37 (d, 3J ϭ 7.8 Hz,
only in stoichiometric amounts. Thus, while these com-
plexes are all capable of decomposing hydrogen peroxide,
the mechanisms through which H2O2 is destroyed seem to
be different. The decomposition of hydrogen peroxide can-
not reasonably be described as a ‘‘catalase-like’’ activity.
One possible reason might be that the chemical structure of
the respective ligands does not allow easy formation of a π-
type radical cation, which plays a central role in the activity
of natural iron-containing catalases.
3
2 H, 22-H, 24-H), 7.82 (dd, J ϭ 7.8 Hz, 1 H, 23-H). Ϫ 13C NMR
Experimental Section
(125 MHz, [D6]DMSO): δ ϭ 46.4 (C-2, C-20), 54.2 (C-10, C-12),
109.9 (C-5, C-17), 115.9 (C-7, C-15), 120.2 (C-22, C-24), 124.3 (C-
9, C-13), 128.6 (C-6, C-16), 130.4 (C-8, C-14), 137.5 (C-23), 147.9
1
Instrumentation: H and 13C NMR spectra: Bruker GRX 500 and
Varian Gemini 200. Ϫ High resolution MS: Fisons Instruments VG
Pro Spec 3000 (70 eV). Ϫ Melting points (uncorrected): Elektro-
thermal 9100. Ϫ IR: Bio-Rad Series FTS 135 FT-IR. Ϫ UV/Vis:
Varian Cary 219. Ϫ Oxygen measurements: Clark-type polaro-
graphic oxygen electrode from Eschweiler and Hansatech (oxygen
electrode unit DW1 and oxygen control box CB1-D3) with DigiS
data acquisition system. Ϫ Elemental analysis: Carlo Erba 1010
CHNSO.
(C-4, C-18), 155.5 (C-1, C21). Ϫ IR (KBr): ν [cmϪ1] ϭ 3298 (NH),
˜
3042 (CHPh), 2843 (CHaliph), 1606 (NH). Ϫ UV (dichloromethane):
λmax [nm] (lg ε) ϭ 251 (4.37), 297 (3.92). Ϫ MS (70 eV): m/z (%) ϭ
330 (66) [Mϩ], 209 (100) [Mϩ Ϫ C7H9N2]. Ϫ HR-MS calcd. for:
C21H22N4: 330.184447; found 330.186833.
N,N-Bis(2-cyanoethyl)-2-methoxybenzylamine: 2-Methoxybenzyl-
amine (21.9 g, 0.16 mol), acrylonitrile (14.5 g, 0.27 mol), and benzo-
quinone (0.50 g, 4.63 mmol) (to avoid formation of polymeric by-
products) were placed in a 250-mL ampoule. After flame-sealing,
the reaction mixture was heated at 140 °C for 6 d. The viscous
product was obtained by distillation (18.5 g, 76.0 mmol, 56%). Ϫ
Materials: Starting materials not mentioned below were commer-
cially available. Catalase (beef liver, 65000 U/mg) and superoxide
dismutase (bovine erythrocytes, 5000 U/mg) were purchased from
Boehringer, Mannheim. The following compounds were prepared
according to published procedures: {2,12-dimethyl-3,7,11,17-tetra-
azabicyclo[11.3.1]heptadeca-1(17),2,11,13,15-pentaene}nickel(II)
1
B.p. 185Ϫ195 °C (0.1 hPa). Ϫ H NMR (200 MHz, CDCl3): δ ϭ
3
3
2.43 (t, J ϭ 6.9 Hz, 4 H, 2 ϫ CH2), 2.81 (t, J ϭ 6.9 Hz, 4 H, 2
ϫ CH2), 3.67 (s, 2 H, CH2), 3.79 (s, 3 H, CH3), 6.90 (m, 2 H, 2 ϫ
HPh), 7.27 (m, 2 H, 2 ϫ HPh). Ϫ 13C NMR (50 MHz, CDCl3): δ ϭ
16.7 (CH2), 49.6 (CH2), 51.1 (CH2), 55.5 (CH3), 110.7 (CPh), 119.1
(CHPh), 120.7 (CN), 125.7 (CHPh), 128.9 (CHPh), 130.6 (CHPh),
157.9 (CHPh). Ϫ C14H17N3O (243.3): calcd. C 69.11, H 7.04, N
17.27; found C 69.25, H 7.13, N 17.14.
perchlorate,[20]
{2,12-dimethyl-3,7,11,17-tetraazabicyclo[11.3.1]-
heptadeca-1(17),13,15-triene}nickel(II) perchlorate,[20] (2R,12S)-
2,12-dimethyl-3,7,11,17-tetraazabicyclo[11.3.1]heptadeca-
1(17),13,15-triene monohydrate (1),[16] pyridine-2,6-dicarbal-
dehyde,[21] 3,7,11,17-tetraazabicyclo[11.3.1]heptadeca-1(17),13,15-
triene (2),[18] 1-chloro-2-nitrobenzene,[22] N,N-bis(2-aminobenzyl)a-
mine,[23] 4-oxa-1,7-heptanediol,[24] 4-oxaheptane-1,7-ditosylate.[24]
N,N-Bis(3-aminopropyl)-2-methoxybenzylamine: A solution of alu-
minium chloride (22.1 g, 0.17 mol) in anhydrous diethyl ether
(200 mL) was added to a suspension of lithium aluminium hydride
(6.34 g, 0.17 mol) in anhydrous diethyl ether (330 mL) under argon.
The mixture was vigorously stirred and a solution of N,N-bis(2-
cyanoethyl)-2-methoxybenzylamine (16.0 g, 65.8 mmol) in anhyd-
rous diethyl ether (200 mL) was added over a period of 2 h. After
completion of addition, the mixture was stirred for 90 h. The reac-
tion mixture was then cooled to 0 °C and quenched with 30% pot-
assium hydroxide (w/v). The yellow ether layer was separated and
the aqueous phase was extracted with diethyl ether (5 ϫ 50 mL).
The combined ethereal extracts were dried with sodium sulfate and
filtered, and the diethyl ether was evaporated. The orange, oily res-
idue was distilled to give a viscous product (9.78 g, 38.9 mmol,
59%). Ϫ B.p. 135Ϫ142 °C (0.1 hPa). Ϫ 1H NMR (200 MHz,
CDCl3): δ ϭ 1.00 (br. s, 4 H, 2 ϫ NH2), 1.52 (m, 4 H, 2 ϫ CH2),
N,N-Bis(2-nitrobenzyl)amine: The literature synthesis[25] was im-
proved as follows. An equimolar amount of gaseous ammonia was
bubbled through an ice-cooled solution of nitrobenzyl chloride
(10.0 g, 58.3 mmol) in ethanol (100 mL). The reaction mixture was
allowed to stand for 6 d at room temperature. The crystalline prod-
uct was filtered off and dissolved in chloroform. Insoluble ammo-
nium chloride was filtered off and chloroform was removed by
evaporation to afford the product as yellow needles (6.01 g,
1
26.8 mmol, 92%). Ϫ M.p. 100 °C (ref.[25] 99Ϫ100 °C). Ϫ H NMR
(200 MHz, CDCl3): δ ϭ 1.98 (s, 1 H, NH), 4.08 (s, 4 H, 2 ϫ CH2),
7.44 (m, 2 H, 2 ϫ CHPh), 7.60 (m, 4 H, 4 ϫ CHPh), 7.95 (m, 2 H,
2 ϫ CHPh). Ϫ 13C NMR (50 MHz, CDCl3): δ ϭ 49.8 (CH2), 124.6
(CHPh), 124.6 (CHPh), 128.5 (CHPh), 131.7 (CHPh), 133.2 (CHPh),
134.4 (CHPh). Ϫ IR (KBr): ν [cmϪ1] ϭ 3247 (NH), 3080 (CHPh),
˜
2909 (CHaliph), 1610 (NH), 1516, 1337 (NO2).
3
3
2.39 (t, J ϭ 6.9 Hz, 4 H, 2 ϫ CH2), 2.61 (t, J ϭ 6.9 Hz, 4 H, 2
ϫ CH2), 3.47 (s, 2 H, CH2), 3.71 (s, 3 H, CH3), 6.85 (m, 4 H, 4 ϫ
HPh). Ϫ 13C NMR (50 MHz, CDCl3): δ ϭ 30.8 (CH2), 40.3 (CH2),
51.4 (CH2), 51.86 (CH2), 54.9 (CH3), 109.8 (CHPh), 119.9 (CHPh),
127.3 (CHPh), 127.6 (CHPh), 129.7 (CHPh), 157.3 (CHPh). Ϫ IR
3,7,19,25-Tetraazatetracyclo[19.3.1.04,9013,18]pentacosa-1(25),4,6,
8,13,15,17,21,23-nonaene (3): A solution of cupric nitrate trihydrate
(1.62 g, 6.70 mmol) in water (20 mL) was added to a stirred solu-
tion of pyridine-2,6-dicarbaldehyde (0.91 g, 6.70 mmol) in ethanol
(20 mL).
A solution of N,N-bis(2-aminobenzyl)amine (1.52 g,
(HATR): ν [cmϪ1] ϭ 3293 (NH), 3063 (CHPh), 2940 (CHaliph), 2835
˜
6.70 mmol) in ethanol (4 mL) was added dropwise to the resulting
pale green solution, with rapid stirring. The reaction mixture was
refluxed for 2.5 h, cooled down in an ice/brine bath, and sodium
tetrahydroborate (0.66 g, 17.5 mmol) was added in small portions.
After completion of addition, the solution was stirred at room tem-
(CHmethyl ether), 1570 (NH), 1241 (COarylalkyl ether). Ϫ C14H25N3O
(251.4): calcd. C 66.89, H 10.02, N 16.72; found C 66.74, H 10.05,
N 16.53.
7-(2-Methoxybenzyl)-3,7,11,17-tetraazabicyclo[11.3.1]heptadeca-
perature for 2 h, heated at 60 °C for 1 h, and finally stirred at room 1(17),13,15-triene (4a): A solution of cupric nitrate trihydrate
Eur. J. Org. Chem. 2001, 3119Ϫ3125 3123

S. Autzen, H.-G. Korth, H. de Groot, R. Sustmann
FULL PAPER
(10.2 g, 42.1 mmol) in water (90.0 mL) was added to a solution of
pyridine-2,6-dicarbaldehyde (5.69 g, 42.1 mmol) in ethanol (90 mL)
ν
˜ [cm
Ϫ1] ϭ 3257, 3179 (NH), 3065 (CHPh), 2976 (CHaliph), 1606
(NH), 1581 (CNPyr). Ϫ C15H26Cl2BF4FeN4 (476.0): calcd. C 37.85,
whilst stirring. To the resulting pale green mixture was added drop- H 5.51, N 11.77; found C 37.56, H 5.70, N 11.51.
wise a solution of N,N-bis(3-aminopropyl)-2-methoxybenzylamine
(10.6 g, 42.1 mmol) in ethanol (18 mL), with rapid stirring. After
completion of addition (30 min), the resulting dark blue solution
was refluxed for 4 h, then stirred at room temperature overnight.
The reaction mixture was cooled to 5 °C in an ice/brine bath and
sodium borohydride (4.00 g, 0.11 mol) was added over 30 min. The
solution was then stirred at room temperature for 1 h, heated at 60
°C for 1 h, and stirred further at room temperature for several
hours. The copper(II) was removed by treating the mixture with
sodium sulfide nonahydrate (23.1 g, 96.0 mmol). The mixture was
stirred at room temperature for 1 h, and heated at 60 °C for 2 h.
The solution was cooled and the copper(II) sulfide removed by
Ferric Complex of 2 [Fe(III)-2]: The procedure used was adapted
from the literature,[16] in anhydrous methanol (total 175 mL) with
an equimolar amount of ligand 2 (1.98 g, 8.46 mmol) and anhyd-
rous ferrous chloride (1.07 g, 8.46 mmol). Tetraethylammonium
tetrafluoroborate (1.84 g, 8.46 mmol) was employed as the counter-
ion source. The ferrous complex of 2 was not isolated but directly
oxidized to Fe(III)-2 (yellow powder, 2.35 g, 5.25 mmol, 62%), de-
1
comp. at 145 °C. Ϫ H NMR (500 MHz, D2O): δ ϭ 2.29 (br. s, 4
H, 5-H, 9-H), 3.20 (br. s, 4 H, 6-H, 8-H), 3.32 (br. s, 4 H, 4-H, 10-
H), 4.56 (br. s, 4 H, 2-H, 12-H), 7.57 (br. s, 2 H, 14-H, 16-H), 8.01
(br. s, 1H 15-H). Ϫ 13C NMR (125 MHz, D2O): δ ϭ 23.4 (C-5, C-
9), 44.6 (C-6, C-8), 45.6 (C-4, C-10), 52.0 (C-2, C-12), 127.3 (C-14,
filtration through Celite. Ethanol was evaporated, the residue ex-
C-16), 142.7 (C-15), 152.8 (C-1, C-13). Ϫ IR (KBr): ν [cmϪ1] ϭ
˜
tracted three times with dichloromethane, and the combined ex-
tracts dried with MgSO4. After removal of the dichloromethane by
evaporation the brown residue was purified by crystallization from
chloroform (several days in a refrigerator) (8.65 g, 24.4 mmol,
3657, 3255 (NH), 3084 (CHPh), 2945 (CHaliph), 1609 (NH). Ϫ UV
(water): λmax [nm] (lg ε) ϭ 254 (3.82), 276Ϫ450 br. sh. Ϫ
C13H22Cl2BF4FeN4 (447.9): calcd. C 34.86, H 4.95, Fe 12.47, N
12.51; found C 35.05, H 5.08, Fe 12.25 (AAS), N 12.42.
1
58%). Ϫ M.p. 269 °C. Ϫ H NMR (500 MHz, CDCl3): δ ϭ 1.52
3
(m, 4 H, 5-H, 9-H), 2.18 (t, J ϭ 6.0 Hz, 4 H, 4-H, 10-H), 2.30 (t,
Ferric Complex of 3 [Fe(III)-3]: This compound was prepared as
described above for Fe(III)-2. Fe(III)-3 was obtained as a green
powder (54%). Ϫ IR (KBr): ν˜ [cmϪ1] ϭ 3630 (NH), 3059 (CHPh),
2931 (CHaliph), 1633 (NH), 1060 (BF4). Ϫ UV (dichloromethane):
3J ϭ 6.0 Hz, 4 H, 6-H, 8-H), 3.32 (s, 2 H, CH2), 3.46 (s, 3 H,
OCH3), 3.67 (s, 4 H, 2-H, 12-H), 6.59 (m, 2 H, 2 ϫ CHPh), 6.80
3
(d, J ϭ 7.3 Hz, 2 H, 14-H, 16-H), 6.97 (m, 2 H, 2 ϫ CHPh), 7.34
(dd, 3J ϭ 7.3 Hz 1 H, 15-H). Ϫ 13C NMR (125 MHz, CDCl3): δ ϭ
27.7 (C-5, C-9), 46.2 (C-4, C-10), 51.8 (CH2), 52.3 (C-6, C-8), 53.5
(C-2, C-12), 55.0 (OCH3), 110.1 (CHPh) 120.0 (CHPh), 120.5 (C-
14, C-16), 127.3 (CHPh), 127.6 (CHPh), 130.5 (CHPh), 136.4 (C-15),
λmax [nm] (lg ε)
ϭ
272 (4.36), 273Ϫ400 br. sh.
Ϫ
C21H22Cl2BF4FeN4 (544.0): calcd. C 46.37, H 4.08, N 10.30; found
C 46.02, H 3.89, N 10.21.
157.8 (CHPh), 159.4 (C-1, C-13). Ϫ IR (KBr): ν [cmϪ1] ϭ 3296
˜
Attempted Preparation of the Ferric Complex of 4a [Fe(III)-4a]: Nu-
merous attempts to prepare the iron(III) complex of 4a were made,
by variation of the reaction conditions as summarized here. Iron(II)
salts: FeCl2, Fe(CH3CO2)2, Fe(ClO4)2·6 DMSO, and Fe(BF4)2 and
(NH), 3060 (CHPh), 2935 (CHaliph), 2835 (CHmethyl ether), 1590
(NH), 1240 (COarylalkyl ether). Ϫ UV (water): λmax [nm] (lg ε) ϭ 261
(3.54), 266 (3.53), 276 sh (3.24). Ϫ HR-MS calcd. for C21H30N4O:
354.241962 found 354.242928.
subsequent oxidation by air to FeIII
, with BF4 , ClO4 ,
Ϫ
Ϫ
B(C6H5)4Ϫ, or PF6Ϫ, respectively, as counter-ion; iron(III) salts:
FeCl3, Fe(ClO4)3·6 H2O, and Fe(ClO4)3·6 DMSO; solvents (dry):
methanol, ethanol, tetrahydrofuran, acetonitrile; bases: triethylam-
ine; concentration ratios of ligand/FeII/III salt ϭ 2:1, 1:1, 1:2, 1:10;
reaction times: 10 min, 1 h, 2 h, 1 d, 1 week; reaction temperatures:
20 °C, 40 °C, reflux; pressure: 1 bar, 10 kbar. In no case could a
well-defined Fe(III)-4a complex be isolated.
7-(2-Hydroxybenzyl)-3,7,11,17-tetraazabicyclo[11.3.1]heptadeca-
1(17),13,15-triene (4b): A solution of BBr3 in dichloromethane (1 ,
5.72 mL, 5.72 mmol) was added dropwise to an ice/brine-cooled
solution of 7-(2-methoxybenzyl)-3,7,11,17-tetraazabicyclo[11.3.1]-
heptadeca-1(17),13,15-triene (340.3 mg, 0.96 mmol) in anhydrous
dichloromethane (8.0 mL). After completion of addition, the reac-
tion was stirred at room temperature for 20 h, water (14.0 mL) was
added dropwise in order to hydrolyze the excess of the reagent, and
the mixture was stirred intensively for 2 h. The aqueous phase was
extracted with dichloromethane (4 mL), brought to pH ϭ 11 with
2.5  NaOH, and extracted with dichloromethane (3 ϫ 4 mL).
These extracts were dried with MgSO4, and the dichloromethane
removed by evaporation to leave a colorless viscous product
(204 mg, 0.60 mmol, 62%). Ϫ 1H NMR (500 MHz, CDCl3): δ ϭ
1.74 (m, 4 H, 5-H, 9-H), 2.43 (t, 3J ϭ 5.8 Hz, 4 H, 4-H, 10-H),
Oxygen Measurements: To monitor the dioxygen evolution initiated
by Fe(III)-1, an Eschweiler M100LCD polarographic Clark-type
oxygen electrode was used. A Hansatech DW1/CB1-D3 electrode
was used for Fe(III)-2 and Fe(III)-3. All solutions were purged with
N2 or Ar before being introduced into the chamber. A magnetic
stirrer and the electrode membranes were placed at the bottom of
the reaction vessel, which was surrounded by a water jacket con-
nected to a thermostated reservoir maintained at 25.0Ϯ0.2 °C. Ϫ
The Eschweiler M100 device was calibrated by flushing the buffer
solution with N2 and O2, respectively, to read the zero and maximal
concentrations of oxygen in the solution. The Hansatech DW1/
CB1-D3 device was calibrated with sodium dithionite to identify
the discrepancy between zero oxygen and the electrical zero.
Na2S2O4 consumes O2 according to the equation Na2S2O4 ϩ O2 ϩ
H2O Ǟ NaHSO4 ϩ NaHSO3. The electrical current generated by
the reduction of oxygen at the cathode is converted into a voltage
output signal. In order to convert the voltage output reading of the
instrument to concentration units (µ), a linear calibration curve
was established with known oxygen concentrations. For this pur-
pose, the reactor was filled to a volume of 1.0 mL with varying
3
2.48 (t, J ϭ 6.8 Hz, 4 H, 6-H, 8-H), 3.50 (s, 2 H, CH2), 3.87 (s, 4
H, 2-H, 12-H), 6.59 (m, 1 H, CHPh), 6.66 (m, CHPh), 6.86 (m, 1
3
H, CHPh), 6.98 (d, J ϭ 7.8 Hz, 2 H, 14-H, 16-H), 7.01 (m, 1 H,
CHPh), 7.52 (dd, 3J ϭ 7.8 Hz, 1 H, 15-H). Ϫ 13C NMR (125 MHz,
CDCl3): δ ϭ 27.1 (C-5, C-9), 45.3 (C-4, C-10), 51.5 (C-6, C-8), 53.4
(C-2, C-12), 58.4 (CH2), 115.8 (CHPh), 118.9 (CHPh), 120.8 (C-14,
C-16), 122.4 (CHPh), 128.4 (CHPh), 128.8 (CHPh), 136.8 (C-15),
156.9 (CHPh), 158.5 (C-1, C-13). Ϫ IR (HATR): ν [cmϪ1] ϭ 3280
˜
(OH), (NH), 3060 (CHPh), 2932 (CHaliph), 1590 (NH). Ϫ MS
(70 eV): m/z (%) ϭ 341 (75) [Mϩ ϩ 1], 247 (21) [Mϩ Ϫ C6H5O],
233 (43) [Mϩ Ϫ C7H7O], 107 (100) [Mϩ Ϫ C13H21N4].
Ferric Complex of 1 [Fe(III)-1]: The complex was prepared accord- H2O2 concentrations (100 µ, 200 µ, 300 µ, 400 µ, and 500 µ)
ing to the method of Busch[16] (yellow powder, 49%). Ϫ IR (KBr): in phosphate buffer. After addition of pure beef catalase (1 µL
3124
Eur. J. Org. Chem. 2001, 3119Ϫ3125

Reactions of Tetraazamacrocyclic FeIII Complexes with Hydrogen Peroxide
FULL PAPER
[10]
I. A. Salem, M. El-Maazawi, A. B. Zaki, Int. J. Chem. Kinet.
2000, 32, 643Ϫ666.
G. Tachiev, J. A. Roth, A. R. Bowers, Int. J. Chem. Kinet. 2000,
32, 24Ϫ35.
T. Okuno, S. Ito, S. Ohba, Y. Nishida, J. Chem. Soc., Dalton
Trans. 1997, 3547Ϫ3551.
R. Than, A. Schrodt, L. Westerheide, R. van Eldik, B. Krebs,
Eur. J. Inorg. Chem. 1999, 1537Ϫ1543.
I. Käpplinger, H. Keutel, E.-G. Jäger, Inorg. Chim. Acta 1999,
291, 190Ϫ206.
A. C. Melnyk, N. K. Kildahl, A. R. Rendina, D. H. Busch, J.
Am. Chem. Soc. 1979, 101, 3232Ϫ3240.
C. J. Cairns, R. A. Heckman, A. C. Melnyk, J. M. Davis, D.
H. Busch, J. Chem. Soc., Dalton Trans. 1987, 2505Ϫ2510.
X. P. Zhang, D. L. Zhang, D. H. Busch, R. van Eldik, J. Chem.
Soc., Dalton Trans. 1999, 2751Ϫ2758.
K. P. Balakrishnan, H. A. A. Omar, P. Moore, N. W. Alcock,
G. A. Pike, J. Chem. Soc., Dalton Trans. 1990, 2965Ϫ2969.
J. D. Rush, Z. Maskos, W. H. Koppenol, in: Methods in En-
zymology, vol. 186 (Ed.: L. Packer), Academic Press, San Di-
ego, 1990, pp. 148Ϫ156.
J. L. Karn, D. H. Busch, Inorg. Chem. 1969, 8, 1149Ϫ1153.
N. W. Alcock, R. G. Kingston, P. K. Moore, C. Pierpoint, J.
Chem. Soc., Dalton Trans. 1984, 1937Ϫ1943.
R. E. Bowman, P. J. Islip, I. M. Lockhart, K. E. Richards, M.
Wright, J. Chem. Soc. 1965, 1080Ϫ1087.
D. S. C. Black, N. E. Rothnie, Aust. J. Chem. 1983, 36,
1141Ϫ1147.
J. S. Bradshaw, J. M. Guynn, S. G. Wood, B. E. Wilson, N. K.
Dalley, R. M. Izatt, Heterocycl. Chem. 1987, 24, 415Ϫ419.
S. Gabriel, R. Jansen, Ber. Dtsch. Chem. Ges. 1891, 24,
3091Ϫ3098.
injection volume), the output signal was correlated with the known
oxygen concentration. Ϫ Oxygen release initiated by complexes
Fe(III)-1ϪFe(III)-3 was carried out similarly, by injection of an ali-
quot of a 1 m stock solution of the complex in buffer [a buffer/
DMSO (99:1) solution was used in the case of Fe(III)-3] to a meas-
ured volume of buffer solution containing the desired amount of
hydrogen peroxide, to give a total volume of 1.00 mL.
[11]
[12]
[13]
[14]
[15]
Acknowledgments
This work was supported by the Deutsche Forschungsgemeinschaft
[16]
(SFB 452).
[17]
[1]
´
´
M. Zamocky, F. Koller, Prog. Biophys. Mol. Biol. 1999, 72,
[18]
[19]
19Ϫ66.
[2]
[3]
P. Gouet, H. M. Jouve, O. Dideberg, J. Mol. Biol. 1995, 249,
933Ϫ954.
B. Halliwell, J. M. Gutteridge, Free Radicals in Biology and
Medicine, 3rd ed., Oxford University Press, Oxford, 1999.
Y. Watanabe, H. Fujii, Struct. Bond., 2000, 97, 61Ϫ89.
R. Weiss, D. Mandon, T. Wolter, A. X. Trautwein, M. Muther,
E. Bill, A. Gold, K. Jayaraj, J. Terner, J. Biol. Inorg. Chem.
1996, 1, 377Ϫ383.
S. R. Doctrow, K. Huffman, C. B. Marcus, W. Musleh, A.
Bruce, M. Baudry, B. Malfroy, in: Antioxidants in Disease Ϫ
Mechanisms and Therapy (Ed.: H. Sies), Academic Press, San
Diego, 1997, pp. 247Ϫ269.
[20]
[21]
[4]
[5]
[22]
[23]
[24]
[25]
[6]
[7]
P. A. MacFaul, D. D. M. Wayner, K. U. Ingold, Acc. Chem.
Res. 1998, 31, 159Ϫ162.
[8]
[9]
C. Walling, Acc. Chem. Res. 1998, 31, 155Ϫ157.
S. Goldstein, D. Meyerstein, Acc. Chem. Res. 1999, 32,
547Ϫ550.
Received January 19, 2001
[O01024]
Eur. J. Org. Chem. 2001, 3119Ϫ3125
3125