Monooxygenation of aromatic compounds by dioxygen with bioinspired systems using non-heme iron catalysts and tetrahydropterins: Comparison with other reducing agents and interesting regioselectivity favouring meta-hydroxylation

Article abstract of DOI:10.1016/j.tet.2004.03.006

Monooxygenation of aromatic compounds by dioxygen in the presence of catalytic amounts of an iron(II) salt and tetrahydropterins as reducing agents occurs with a regioselectivity favouring meta-hydroxylation of arenes bearing an electron-donating substituent, such as anisole, phenetole, toluene, and ethylbenzene. Comparison of similar systems using various reducing agents showed that only tetrahydropterins and ascorbate led to such a major meta-hydroxylation. The tetrahydropterin- and ascorbate-dependent systems should be useful for the preparation of meta-hydroxylated metabolites of aromatic drugs, as shown here in the case of diclofenac.

Full text of DOI:10.1016/j.tet.2004.03.006

Tetrahedron 60 (2004) 3855–3862
Monooxygenation of aromatic compounds by dioxygen with
bioinspired systems using non-heme iron catalysts and
tetrahydropterins: comparison with other reducing agents and
interesting regioselectivity favouring meta-hydroxylation
*
Delphine Mathieu, Jean Franc¸ois Bartoli, Pierrette Battioni and Daniel Mansuy
´ `
UMR 8601, Universite Paris V, 45 rue des Saints-Peres, 75270 Paris Cedex 06, France
Received 25 November 2003; revised 10 February 2004; accepted 2 March 2004
Abstract—Monooxygenation of aromatic compounds by dioxygen in the presence of catalytic amounts of an iron(II) salt and
tetrahydropterins as reducing agents occurs with a regioselectivity favouring meta-hydroxylation of arenes bearing an electron-donating
substituent, such as anisole, phenetole, toluene, and ethylbenzene. Comparison of similar systems using various reducing agents showed that
only tetrahydropterins and ascorbate led to such a major meta-hydroxylation. The tetrahydropterin- and ascorbate-dependent systems should
be useful for the preparation of meta-hydroxylated metabolites of aromatic drugs, as shown here in the case of diclofenac.
q 2004 Elsevier Ltd. All rights reserved.
1. Introduction
site.17 These recent data led us to investigate the possible
specific roles played by tetrahydropterins in monooxygen-
ase model systems using non-heme iron complexes as
catalysts, O2 as a source of oxygen atoms, and a reducing
agent. A long time ago, preliminary experiments had shown
that a system based on H4B in the presence of catalytic
amounts of Fe(SO4)2(NH4)2 and EDTA was able to
hydroxylate the aromatic ring of phenylalanine.18 Very
recently, we have compared the properties of similar
systems using different iron catalysts and reducing agents
towards the hydroxylation of aromatic compounds. Here,
we report that some of these systems using tetrahydropterins
are efficient for aromatic hydroxylations, and show an
unusual regioselectivity favouring the meta-hydroxylation
of aromatic compounds bearing an electron-donating
substituent.
Efficient systems based on iron porphyrins or non-heme iron
complexes have been shown to mimic alkene epoxidation
and alkane hydroxylation by cytochrome P450 and non-
heme dependent monooxygenases.1–5 Cytochromes P450
and non-heme iron enzymes such as tetrahydrobiopterin
(H4B)-dependent monooxygenases also catalyse the selec-
tive hydroxylation of aromatic compounds.1–6 However,
few model systems based on iron catalysts have been
described so far for selective and efficient aromatic
hydroxylation.1–5 The most efficient ones use H2O2 as
oxygen atom donor in the presence of either an iron
porphyrin bearing electron-withdrawing b-substituents,7–9
or a polyazamacrocyclic iron complex, Fe(tris-[N-(2-
pyridylmethyl)-2-aminoethyl]amine)(ClO4)2, Fe(TPAA).10
Many systems using a FeII salt, O2 as oxygen atom donor,
and a reducing agent, such as ascorbate in the Udenfriend
system,11 have also been shown to be of interest for the
hydroxylation of aromatic compounds (see for instance Ref.
12–15).
2. Results and discussion
2.1. Hydroxylation of anisole by a FeII/EDTA/O2/
tetrahydropterin system
Tetrahydrobiopterin, H4B, is a necessary cofactor in non-
heme iron aromatic aminoacid monooxygenases,6 and in
NO synthases.16 In the latter enzymes, it plays a key role by
transferring an electron to the heme inside the NOS active
Reaction of anisole with an aerobic phosphate buffer pH 7.4
containing H4B and catalytic amounts of Fe(SO4)2(NH4)2
and EDTA, for 3 h at 20 8C under conditions similar to those
previously used by Viscontini et al. in the case of
phenylalanine,18 led to a mixture of ortho-, meta- and
para-methoxyphenols in a very unusual9,10 36:46:18 ratio,
and to minor amounts of phenol which should come from an
Keywords: Aromatic hydroxylation; Tetrahydrobiopterin; Iron salt;
Diclofenac; Ascorbate.
*
Corresponding author. Tel.: þ33142862187; fax: þ33142868387;
e-mail address: daniel.mansuy@univ-paris5.fr
0040–4020/$ - see front matter q 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tet.2004.03.006

3856
Table 1. Iron-catalyzed hydroxylation of anisole by O2 and diMeH4Pa
Aromatic hydroxylation
D. Mathieu et al. / Tetrahedron 60 (2004) 3855–3862
Other reaction
Turnovers/catalyst
Regioselectivity (%)
Phenol (turnovers/catalyst)
o-OH
m-OH
p-OH
Complete systema
2FeII(SO4)2(NH4)2
2diMeH4P
2Na borohydride
2diMeH4P – Na borohydride
2O2
5.8
,0.2
0.3
1.2
,0.2
,0.2
30
—
—
35
—
—
43
—
—
43
—
—
27
—
—
22
—
—
0.8
,0.2
0.2
0.3
,0.2
,0.2
a
Complete system: 1 mL aerobic 0.1 M phosphate buffer pH 7.4, containing Fe(SO4)2(NH4)2, EDTA and diMeH4P (molar ratio¼1:8:16), and 1 mL anisole
are vigorously stirred for 2 h at 20 8C; [iron catalyst]¼1 mM. Sodium borohydride (10 equiv. relative to diMeH4P) was progressively added during the first
hour to the reaction medium in order to regenerate diMeH4P, at least in part. o-OH, m-OH and p-OH are used for ortho-, meta- and para-methoxyphenol.
Scheme 1.
oxidative dealkylation of anisole. Identification of the
products was performed by GC and confirmed by
GC–MS. Similar results (Table 1) were obtained using
the more accessible 6,7-dimethyltetrahydropterin, diMeH4P
(Scheme 1), instead of H4B. However, replacement of
H4B with dihydrobiopterin, H2B, led to an inactive system.
In their H4B-containing system, Viscontini et al.18 have
described the beneficial effects on the yields of the addition
of a co-reductant, sodium borohydride, which was supposed
to regenerate H4B, at least in part, during the reaction.
Table 1 shows that the addition of an excess of sodium
borohydride to the FeII/EDTA/O2/diMeH4P system led to a
marked increase of the aromatic hydroxylation yield (5.8
turnovers of the iron catalyst instead of 1.2). The absence of
dioxygen (identical reaction under anaerobic conditions), or
of the catalyst (Fe(SO4)2(NH4)2), or of the reducing agents
(diMeH4Pþsodium borohydride), led to an almost inactive
system (Table 1). Moreover, the absence of diMeH4P alone
led to a dramatic decrease of the yields, indicating that
sodium borohydride is almost unable to act as an efficient
reducing agent by itself and that its beneficial effects in the
complete system is due to its regeneration of diMeH4P
during the reaction.
expressed relative to starting diMeH4P, simply because
sodium borohydride rapidly reacts with water and only a
small, non measurable part of it is used to regenerate
diMeH4P.
Kinetic experiments on the oxidation of anisole by the
FeII/EDTA/O2/diMeH4P/NaBH4 system showed that the
reaction was almost complete after 2 h, with an initial rate of
methoxyphenols formation of 3.5 turnovers of the iron
catalyst per hour (Fig. 1). Stopping of the reaction was
presumably due to the consumption of the reducing agents,
as a further addition of diMeH4P and sodium borohydride
led to similar yields and rates. Figure 1 shows that the
regioselectivity of anisole hydroxylation varies as a function
of the reaction time. At the beginning of the reaction (,1 h),
the relative importance of meta-hydroxylation is at its
maximum level, with an ortho:meta:para ratio of about
25:50:25, whereas this ratio tends towards 34:36:30 at the
end of the reaction. This could mean that meta hydroxyl-
ation of anisole is favoured in the presence of the high
concentrations of reductant that occur at the beginning of
the reaction, and that could keep the Fe(II) concentration
high.
In Table 1 and the following tables, yields of hydroxylated
products are generally expressed in turnovers of the iron
catalyst (mol of product formed during the reaction per mol
of catalyst). This parameter gives an idea of the catalytic
efficiency of the system. From time to time, yields are also
expressed in mol of product per mol of starting reducing
agent (in %). This second parameter gives the number of
electrons used by the system for dioxygen activation and
hydroxylation of the aromatic substrates. It is an index of the
decoupling between electron consumption by the system
and hydroxylation of the substrate. In most bioinspired
systems using an iron catalyst, O2 and a reducing agent, this
decoupling is large and yields based on the reducing agent
are low.15 In the system using both diMeH4P and sodium
borohydride (in order to regenerate diMeH4P), yields are
Figure 1. Kinetic study of the formation of o-, m- and p-methoxyphenols
upon oxidation of anisole by the FeII/EDTA/diMeH4P/NaBH4 system
(conditions of Table 1).

D. Mathieu et al. / Tetrahedron 60 (2004) 3855–3862
3857
Table 2. Hydroxylation of anisole by O2 and diMeH4P in the presence of Fe(SO4)2(NH4)2 and various chelatesa
Chelate
Aromatic hydroxylation
Total yield/diMeH4P (%)
Total turnovers/catalyst
Regioselectivity (%)
o-OH
m-OH
p-OH
Without EDTA
EDTA
EGTA
EDDA
DTPA
HETA
,0.2
5.8
0.2
3.2
1.5
,1
37
1
20
9.5
18
—
30
45
30
47
33
—
43
45
45
32
39
—
27
10
25
21
28
2.9
a
Conditions as in the complete system described in Table 1.
Figure 2. Formula of the various chelates used in this study.
2.2. Studies of the reaction parameters
sulfonato-porphyrin]21
and
microperoxidase,
MP11Fe(III),22 failed to yield significant amounts of
methoxyphenols.
Table 2 shows the importance of the EDTA ligand on
anisole hydroxylation by the above mentioned system. In
the absence of EDTA, the system was almost unable to
catalyze this reaction. Replacement of EDTA with
analogous polydentate ligands, such as EGTA, EDDA,
DTPA and HETA† (Fig. 2) led to a marked decrease of the
methoxyphenols yields (0.2–3.2 turnovers/iron instead of
5.8 in the case of EDTA). This decrease of the yields was
accompanied by changes of the regioselectivity which was
still in favour of meta-methoxyphenol except in the case of
DTPA. The distances between the coordinating groups of
EDTA seem to play a crucial role, since replacement
of EDTA with EGTA, which has a longer chain led to a
dramatic decrease of the yields (Table 2).
Moreover, a study of the effects of the concentration of the
iron salt showed us that the yield of meta-methoxyphenol
increased with the FeII salt concentration up to 0.5–1 mM
(Table 3). For concentrations higher than 1 mM, the yield of
anisole hydroxylation dramatically decreased. Therefore, a
FeII concentration of 1 mM was used in the following
studies.
Table 3. Effects of FeII concentration on iron-catalyzed hydroxylation of
anisole by O2 and diMeH4 Pa
[FeII] (M)
Aromatic hydroxylation
Total turnovers/catalyst Regioselectivity (%)
A study of the effects of the FeII/EDTA molar ratio on
anisole hydroxylation showed that the best compromise for
reaching good hydroxylation yields and a regioselectivity in
favour of meta-hydroxylation was obtained with a 1:10 ratio
(data not shown).
o-OH
m-OH
p-OH
4£1025
5£1024
1023
0.8
6.9
5.8
0.7
0.2
80
39
30
29
50
10
30
43
43
25
10
31
27
28
25
2£1023
4£1023
Replacement of Fe(SO4)2(NH4)2 with the various non-heme
iron complexes, [(TPAA)FeII](ClO4)2,10 (L52FeCl)PF6 19 and
[Fe(TPA)(CH3CN)2](ClO4)2,20 or with iron porphyrin
complexes, such as Fe(III)[tetra-pentafluorophenyl-b-tetra-
a
Conditions as in the complete system described in Table 1; molar ratio
1:8:16:160 for FeII/EDTA/diMeH4P/sodium borohydride was conserved.
Finally, a study of the effects of pH on the hydroxylation of
anisole showed that the total yield of methoxyphenols reach
a maximum value at pH 7.4 (5.8 turnovers) (Table 4).
However, the relative importance of meta-hydroxylation
was enhanced when pH was increased from 4.5 up to 9, and
was maximum at pH 9 (52% of total methoxyphenols). The
†
EDTA, EGTA, EDDA, DTPA and HETA correspond to ethylenediamine
tetraacetic acid, ethylenebis(oxyethylenenitrilo)-tetraacetic acid,
ethylenediamine-N,N0-diacetic acid, diethylenetriamine pentaacetic acid
and N(2-hydroxyethyl)ethylenediaminetriacetic acid respectively (see
Fig. 2).

3858
D. Mathieu et al. / Tetrahedron 60 (2004) 3855–3862
Table 4. Effects of pH on iron(II)-catalyzed hydroxylation of anisole by O2
and diMeH4 Pa
described previously9,10 almost exclusively led to ortho-
and para-hydroxylation of anisole and ethoxybenzene
(meta-hydroxylation ,2% of total aromatic hydroxylation),
as expected for systems involving electrophilic hydroxyl-
ating species. With the present system, hydroxylation of the
aromatic ring of toluene and ethylbenzene also occurred
mainly at the meta position (50% of total aromatic
hydroxylation, Table 5), whereas the previous iron cata-
lyst/H2O2 systems7–10 gave much lower levels of meta-
hydroxylation for these two substrates (between 11 and
25%). This unusual regioselectivity of the diMeH4P-
dependent system was particularly well illustrated in the
case of diclofenac (Scheme 2). This anti-inflammatory drug
is metabolized by human cytochromes P450 with major
formation of 40-hydroxydiclofenac and minor formation of
5-hydroxydiclofenac, which are both derived from
hydroxylations at the para positions of the strong electron-
donating NH group.23,24 Iron porphyrin–H2O2 (or
tBuOOH) systems mainly hydroxylate diclofenac at
position 5, 40-hydroxydiclofenac being formed in small
amounts.23 In contrast, the diMeH4P-dependent system led
to significant amounts of 30-hydroxydiclofenac, which
results from meta-hydroxylation of the dichlorophenyl
ring of diclofenac (32% of total aromatic hydroxylation,
Table 5). This 30-hydroxy metabolite, which was only
formed in trace amounts with cytochromes P45024 and iron
porphyrin model systems,23 was produced by the diMeH4P-
dependent system in amounts comparable to those of 40- and
5-hydroxydiclofenac (Table 5).
pH
Aromatic hydroxylation
Total turnover/catalyst Regioselectivity (%)
o-OH
m-OH
p-OH
4.5
6
7.4
8.2
9
1.1
5.1
5.8
3.4
2.5
1.8
54
44
30
35
29
39
31
34
43
42
52
39
15
22
27
23
19
22
11.8
a
Conditions as in the complete system described in Table 1, except for the
buffer which was adapted to each pH (see Section 4).
following experiments were performed at pH 7.4 that was a
good compromise for good yields and regioselectivity.
2.3. Generalization to various substrates
The Fe(SO4)2(NH4)2/EDTA/O2/diMeH4P system was also
found to hydroxylate other aromatic compounds, such as
benzene itself, with formation of the corresponding phenols
(total turnovers of catalyst for aromatic hydroxylation
between 1 and 5.8, and yields based on diMeH4P between
6.5 and 37%) (Table 5). The most distinctive and surprising
characteristic of this system is the unusual regioselectivity
of its hydroxylation of aromatic compounds bearing an
electron-donating substituent. Thus, hydroxylation of
anisole mainly occurred at the meta position (43% of total
aromatic hydroxylation), and meta-hydroxylation of
ethoxybenzene reached 38% of total aromatic hydroxyl-
ation. In contrast, the efficient iron catalyst/H2O2 systems
2.4. Effects of the replacement of diMeH4P by other
reducing agents
In order to better understand the origin of the important
Table 5. Regioselectivity of the hydroxylation of aromatic compounds by O2 and diMeH4P in the presence of Fe(SO4)2(NH4)2 and EDTAa
Substrate
Aromatic hydroxylation
Total turnovers/catalyst Total yield (% based on diMeH4P)b Regioselectivity (%)
o-OH m-OH p-OH
Other reactions yields (% based on diMeH4P)b,c
Anisole
Ethoxybenzene
Toluene
Ethylbenzene
Benzene
Diclofenac
5.8
4
2.8
1
5
2.5
37
25
17.5
6.5
31
30
33
40
40
43
38
50
50
27
29
10
10
5 (phenol)
4 (phenol)
3 (PhCH2OH)
74 (PhCOCH3)þ17 (PhCHOHCH3)
15.5
32 (30-OH): 36 (40-
OH): 32 (5-OH)
a
Conditions as in the complete system described in Table 1. In the case of diclofenac, saturating amounts of diclofenac sodium salt were dissolved in the
phosphate buffer medium.
Yields based on starting diMeH4P; one turnover of the catalyst roughly corresponds to 6% yield.
Reactions occurring in competition with aromatic hydroxylation are oxidative dealkylation for anisole and ethoxybenzene, and benzylic hydroxylation in the
case of toluene and ethylbenzene.
b
c
Scheme 2.

D. Mathieu et al. / Tetrahedron 60 (2004) 3855–3862
3859
a
Table 6. Effects of the nature of the reducing agent on iron-catalyzed hydroxylation of anisole by O2
Reducing agent
Aromatic hydroxylation
Total turnovers/catalyst
Total Yieldb/reductant (% based)
Regioselectivity (%)
o-OH
m-OH
p-OH
DiMeH4PþNaBH4
DiMeH4P
H4B
H4F
Ascorbate
Catechol
Hydroquinone
Trimethylhydroquinone
Thiophenol
5.8
1.2
0.9
1.5
6.4
0.7
0.3
2.9
,0.2
,0.2
0.9
,0.2
,0.2
,0.2
37 (3.3)
7.5
5.5
9
40
4.5
2
17.5
,1
,1
5.5
,1
,1
,1
30
35
36
39
31
44
58
39
—
—
36
—
—
—
43
43
46
33
40
—
—
22
—
—
18
—
—
—
27
22
18
28
29
56
42
39
—
—
46
—
—
—
a-Naphthol
Methylhydrazine
2-Mercaptobenzoic acid
Hydrazobenzene
NADH
a
Conditions as in the complete system described in Table 1, but without addition of sodium borohydride (except as specified for the first line of the table). Iron
catalyst/reductant molar ratio¼1/16.
b
Yields are based on the starting reducing agent; in the case of diMeH4PþNa borohydride, an excess of sodium borohydride was also used and yields are
expressed on the basis of starting diMeH4P. Yield in parentheses is based on starting diMeH4Pþsodium borohydride.
meta-hydroxylation observed with the FeII/EDTA/O2/
spectrometry coupled to gaz chromatography. When the
same reaction was performed with 16O2 in H218O (98%
enriched), the methoxyphenol products contained less than
1% 18O. These data clearly show that the oxygen atom
incorporated into anisole almost exclusively came from O2
(Scheme 3), whatever the position, ortho, meta or para, of
the hydroxyl group.
diMeH4P system, we have replaced diMeH4P by various
reductants. Some of them have been previously used in
oxidizing systems based on an iron catalyst, O2 and a
reducing agent. This is the case of catechol,25 hydro-
quinone,26 trimethylhydroquinone,27 2-mercaptobenzoic
acid,28 hydrazobenzene29 and ascorbic acid.11 Table 6
shows that the regioselectivities the most in favour of meta-
hydroxylation of anisole were obtained with tetrahydro-
pterins diMeH4P and H4B, and with ascorbate. Only these
three reducing agents led to a major formation of meta-
methoxyphenol. Tetrahydrofolate, H4F (Scheme 1), another
natural tetrahydropterin cofactor, gave results similar to
H4B and diMeH4P, however with lower relative amounts of
meta-methoxyphenol. Trimethylhydroquinone and methyl-
hydrazine also led to significant amounts of meta-methoxy-
phenol but as a minor product when compared to its ortho
and para isomers. Catechol and hydroquinone exclusively
led to ortho- and para-methoxyphenol. All the other tested
reducing agents, thiophenol, a-naphthol, 2-mercapto-
benzoic acid, hydrazobenzene and NADH failed to produce
methoxyphenols in significant amounts. These data clearly
showed that the structure of the reducing agent is a key
factor in aromatic hydroxylations by the FeII/EDTA/O2/
reductant systems. Two reducing agents emerge from this
comparison, diMeH4P (or H4B) and ascorbate; they are
much superior to the others both in terms of catalytic
efficiency for anisole hydroxylation, with about six turn-
overs of the catalyst, and in terms of regioselectivity in
favour of meta-hydroxylation (43 and 40% of total aromatic
hydroxylation).
Scheme 3.
2.5.2. Possible role of H2O2. All the iron/O2/reductant
model systems of monooxygenases led to a partial
decoupling between the consumption of electrons from the
reducing agent and the monooxygenase reaction, which
results in the formation of products coming from dioxygen
reduction such as H2O2. In order to evaluate the role of
H2O2 possibly produced during the studied reactions,
oxidation of anisole by the FeII/EDTA/O2/diMeH4P system
was performed in the presence of catalase. No significant
difference was noted in the yields observed with or without
catalase, implying that H2O2 does not play an important role
in these reactions.
2.6. Discussion
Our data show that systems based on tetrahydropterins as
reducing agents, in the presence of a catalytic amount of an
iron(II) salt and EDTA, which are closely related to the
system previously used by Viscontini et al. for
the hydroxylation of phenylalanine,18 also catalyse the
hydroxylation of many aromatic compounds such as
benzene, anisole, ethoxybenzene, toluene, ethylbenzene
and diclofenac. The main characteristic of these systems is
their regioselectivity in favour of the meta-hydroxylation of
2.5. Complementary mechanistic experiments
2.5.1. Origin of the oxygen atom inserted into substrates.
Oxidation of anisole by 18O2 (containing 98% 18O) and the
FeII/EDTA/diMeH4P system under conditions similar to
those of Table 1 (see Section 4) led to ortho-, meta- and
para-methoxyphenols bearing a phenolic oxygen atom
containing 98^1% of 18O isotope, as shown by mass

3860
D. Mathieu et al. / Tetrahedron 60 (2004) 3855–3862
aromatic compounds bearing an electron-donating sub-
stituent. A comparative study of many systems using other
reducing agents, under identical conditions, clearly showed
that only diMeH4P (or other tetrahydropterins) and
ascorbate led to such a regioselectivity in favour of meta-
hydroxylation. This suggests that a reinvestigation of meta-
hydroxylation by Udenfriend-type systems (FeII/O2/
ascorbate) should be very much interesting. In fact, there
are few precedents of such iron-catalyzed meta-mono-
oxygenations of aromatic compounds bearing an electron-
donating substituent in the literature (for reviews, see Ref.
12–14). Most of the previously reported iron-catalyzed
oxidations of such aromatic compounds, using either
H2O2,7 – 10,13,15 or O2 and a reductant,12 – 15,30 led to
preferential ortho- and para-hydroxylations. Many reports
on systems related to the one originally described by
Udenfriend et al., and using O2 and ascorbate, have been
published; however very few mentioned meta-hydroxyl-
ation. Hamilton et al. have used a FeClO4/EDTA/O2/
ascorbate system and found at best 18% of meta-
methoxyphenol (relative to total aromatic hydroxylation)
in the hydroxylation of anisole.31 At the same period,
Norman and Lindsay Smith13 have used a similar system
and reported regioselectivities for anisole and toluene
involving 22 and 25% of meta-hydroxylated compound,
respectively. More data have been published on meta-
hydroxylation of aromatic compounds bearing an electron-
donating substituent by another class of systems using FeII
or other reduced metal ions and dioxygen without any
reducing agent.13,14,32–34 These systems are not catalytic
and yields are expressed relative to the metal ion. In the
following, we will only consider the systems using a FeII
salt13,32 – 34 for sake of comparison with our results
exclusively based on FeII catalysts. It has been noticed
that the relative amounts of the meta-hydroxylated products
increased when the iron concentration increased, and the
regioselectivity the most in favour of meta-hydroxylation of
anisole observed by Dearden et al.34 corresponded to a
39:50:11 o:m:p methoxyphenols ratio. These authors also
mentioned that when they added N-benzyl-1,4-dihydro-
nicotinamide as a reducing agent to their system, the meta-
hydroxylated product almost completely disappeared.34
reported in some articles.13,34 In that regard, it is noteworthy
that polymetallic iron complexes are formed upon reaction
of FeII salts with EDTA,35 and that some diiron complexes
are able to catalyze the oxidation of toluene to meta-cresol,
although in very different conditions.36 The reasons why
only two of the tested reducing agents, diMeH4P (or H4B)
and ascorbate, lead to a major meta-hydroxylation of anisole
(Table 6), remain to be determined.
3. Conclusion
Whatever its mechanism may be, the iron/EDTA/
tetrahydropterin/O2 system is interesting because of its
unusual regioselectivity favouring the incorporation of an
oxygen atom from O2 at the meta-position of aromatic
compounds bearing an electron-donating substituent. Our
results show that this is true for many of such aromatic
compounds. They also show that only two reducing agents,
diMeH4P and ascorbate, used in the various
FeII/EDTA/reductant/O2 systems that we have compared,
lead to this major meta-hydroxylation of anisole. Interest-
ingly, the two corresponding systems also lead to the most
efficient catalysis of anisole hydroxylation (Table 6).
Since a fast access to small amounts of various hydroxylated
derivatives of drugs is required for identification of drug
metabolites and for a first evaluation of the pharmacological
properties of these metabolites, the aforementioned
diMeH4P- and ascorbate-dependent systems should be
useful for the preparation of meta-hydroxylated metabolites
of aromatic drugs, as illustrated above in the case of
diclofenac.
4. Experimental
4.1. Reactants
All the reactants and products were commercially available
and purchased from Acros, Sigma, Aldrich and Alfa Aesar
(iron salt). They were used without further purification
except for the aromatic compounds that were eluted on a
small alumina column prior to use. H128O (98% enriched in
18O) and 18O2 (98% enriched in 18O) were purchased from
EURISO-TOP (Saclay, France).
The mechanisms of such meta-monooxygenations of
aromatic compounds are presently almost unknown. The
involvement of electrophilic species such as FeIII–OOH or
high-valent iron–oxo intermediates, that are usually
proposed in iron complex/H2O2 systems, is unlikely,
because these systems have been shown to mainly lead to
ortho- and para-hydroxylated products.7–10,13,15 Meta-
hydroxylation does not seem to be due to the zOH radical,
because the regioselectivity observed for anisole hydroxyl-
ation by zOH appears to correspond to a 84:0:16 molar ratio
of the o-, m-, p-methoxyphenols.13
4.2. General procedure
Hydroxylation of aromatic substrates was performed at
room temperature in vials equipped with a magnetic stirrer.
The organic phase consisting of 1 mL of aromatic substrate
(anisole, phenetole, toluene, ethylbenzene or benzene) was
added to 1 mL of 0.1 M phosphate buffer, pH 7.4, contain-
ing Fe(SO4)2(NH4)2·6H2O (1023 M, 1026 mol, 0.4 mg),
EDTA (8£1023 M, 8£1026 mol, 2.3 mg), diMeH4P
(1.6£1022 M, 1.6£1025 mol, 3.8 mg) and NaBH4
(1.6£1024 mol, 6.4 mg), the latter being progressively
added in one hour. When other reductants than diMeH4P
were used, 1.6£1025 mol was also introduced, but no
sodium borohydride was added. After 2–3 h stirring at
20 8C, an internal standard (PhCOCH3 or PhI,
The involvement of a polymetallic complex, containing at
least two iron atoms, Fe–O and/or Fe–OO moieties, and the
reducing agent, that would first attack the electronically
favoured para (or ortho) position of the aromatic substrate
and then its meta-position, would be at least partly in
agreement with previous proposals from the literature.13,34
This would also fit with the increase of meta-hydroxylation
upon increasing of the iron concentration, as previously

D. Mathieu et al. / Tetrahedron 60 (2004) 3855–3862
3861
1.6£1025 mol) was added and the reaction mixture was
8)/[acetonitrile/water (90/10)], gradient 20% up to 50% in
27 min) was delivered at a rate of 0.6 mL/min. Monitoring
of the column effluent was performed with a detector at 270
and 280 nm. The re0action products were compared to
authentic samples of 3 -, 40- and 5-hydroxydiclofenac kindly
provided by Ciba-Geigy (Basel, Switzerland).
analyzed by gas chromatography.
A strictly similar procedure was followed when pH buffer
was modified. In those cases, we used an acetate buffer
adjusted to pH¼4.5, a citric acid/Na2HPO4 buffer pH¼6, a
trismabase/HCl buffer pH¼8.2, a glycine/NaOH buffer
pH¼9 and a Na2HPO4/NaOH buffer pH¼11.8.
Reactions under anaerobic conditions were done by
‘freeze–thaw cycles’ of a vial containing all the reactants
except the reducing agent, and of a second vial containing
the reducing agent solution under argon. The content of the
first vial was then transferred onto the reductant solution
under argon.
References and notes
1. Meunier, B.; Robert, A.; Pratviel, G.; Bernadou, J. The
porphyrin handbook; Kadish, K. M., Smith, K. M., Guilard,
R., Eds.; Academic Press: New York, 1999; Vol. 4, p 119.
2. Mansuy, D. Coord. Chem. Rev. 1993, 125, 129.
3. Dolphin, D.; Traylor, T. G.; Xie, L. Acc. Chem. Res. 1997, 30,
251.
4.3. Hydroxylation of diclofenac
4. Costas, M.; Chen, K.; Que, L. Coord. Chem. Rev. 2000, 200–
202, 517.
The aqueous phase (1 mL) contained Fe(SO4)2(NH4)2·6H2O
(1023 M), and EDTA (8£1023 M), and 100 mg of the
sodium salt of diclofenac were added to the solution. After
2–3 h stirring at 20 8C, samples were centrifuged to
precipitate excess diclofenac, and products were analyzed
by reverse phase HPLC, as described below.
´
´
5. Menage, S.; Galey, J.-B.; Dumats, J.; Hussler, G.; Seite, M.;
Gautier Luneau, I.; Chottard, G.; Fontecave, M. J. Am. Chem.
Soc. 1998, 120, 13370.
6. Kappock, T. J.; Caradonna, J. P. Chem. Rev. 1996, 96, 2659.
7. Tsuchiya, N.; Seno, M. Chem. Lett. 1989, 263.
8. Carrier, M-N.; Sheer, C.; Gouvine, P.; Bartoli, J. F.; Battioni,
P.; Mansuy, D. Tetrahedron Lett. 1990, 31, 6645.
9. Bartoli, J. F.; Le Barch-Ozette, K.; Palacio, M.; Battioni, P.;
Mansuy, D. J. Chem. Soc., Chem. Commun. 2001, 1718.
10. Bartoli, J. F.; Lambert, F.; Morgenstern-Badarau, I.; Battioni,
P.; Mansuy, D. C.R. Chimie 2002, 5, 263.
4.4. Oxidation of anisole using 18O2
500 mL of 0.1 M phosphate buffer, containing
7.0£1027 mol of Fe(SO4)2(NH4)2, 5.6£1026 mol of
EDTA, and 1 mL of anisole were deaerated by three
freeze–thaw cycles. 1.1£1025 mol of diMeH4P were
dissolved in 200 mL of 0.1 M phosphate buffer (pH 7.4)
and deaerated by the same procedure. 18O2 was introduced
in the main vial and diMeH4P was added using a
deoxygenated syringe. After 2–3 h stirring, products were
analysed by GC–MS as described below.
11. Udenfriend, S.; Clark, C. T.; Axelrod, J.; Brodie, B. B. J. Biol.
Chem. 1954, 208, 731.
12. Hamilton, G. A. In Molecular mechanisms of oxygen
activation; Hayaishi, O., Ed.; Academic: New York, 1974;
p 405.
13. Norman, R. O. C.; Lindsay Smith, J. R. Oxidases and related
redox systems; King, T. E., Mason, H. S., Morrison, M., Eds.;
Wiley: Amherst, 1965; Vol. 1, p 131.
4.5. Oxidation of anisole using H128O
14. Ullrich, V.; Staudinger, Hj. Biological and chemical aspects of
oxygenases: proceedings; Bloch, K. E., Hayaishi, O., Eds.;
Maruzen Scientific: Tokyo, 1966; Vol. 1, p 235.
15. Metal-catalyzed oxidations of organic compounds; Sheldon,
R. A., Kochi, J. K., Eds.; Academic Press: New York, 1981.
16. Pfeiffer, S.; Mayer, B.; Hemmens, B. Angew. Chem. Int. Ed.
1999, 38, 1714.
200 mL of 0.1 M phosphate buffer (pH 7.4) containing
Fe(SO4)2(NH4)2 (1023 M, 2£1027 mol) and EDTA
(8£1023 M, 1.6£1026 mol) were lyophilized and re-dis-
solved in 200 mL of H128O. 500 mL of anisole was added in
the vial. diMeH4P (3.2£1026 mol) was introduced in the
solid state, followed by NaBH4 (1.6£1021 M,
3.2£1025 mol), added in four times. After 2–3 h, products
were analysed by GC–MS as in 18O2 experiments.
17. Hurschman, A. R.; Krebs, C.; Edmondson, D. E.; Huynh,
B. H.; Marletta, M. A. Biochemistry 1999, 38, 15689.
18. Viscontini, M.; Mattern, G. Helvet. Chim. Acta 1970, 53, 372.
19. Bernal, I.; Jensen, I. M.; Jensen, K. B.; McKenzie, C. J.;
Toftlund, H.; Tuchagues, J. P. J. Chem. Soc. Dalton Trans.
1995, 3667.
4.6. Product analysis and identification
GC analyses were done using either a packed 5% FFAP
(polar) column for anisole and benzene, or a capillary BP20
(polar) column for toluene and ethylbenzene, with detection
with a flame ionization detector (FID). The products formed
were analyzed by comparison of their retention time with
those of authentic samples and by gas chromatography-mass
spectrometry analysis using a Hewlett-Packard 5890 Series
II GC coupled with a HP5972 mass selective detector.
20. Zang, Y.; Kim, J.; Dong, Y.; Wilkinson, E. C.; Appelman,
E. H.; Que, L., Jr. J. Am. Chem. Soc. 1997, 119, 4197.
21. Artaud, I.; Ben-Aziza, K.; Mansuy, D. J. Org. Chem. 1993, 58,
3373.
22. Stotter, D. A.; Thomas, R. D.; Wilson, M. T. Bioinorg. Chem.
1977, 7, 87.
23. Mancy, A.; Antignac, M.; Minoletti, C.; Dijols, S.; Mouries,
V.; Duong, N. T.; Battioni, P.; Dansette, P. M.; Mansuy, D.
Biochemistry 1999, 38, 14264.
HPLC analyses of diclofenac metabolites were done using a
Chromatem 380 apparatus. Supernatant aliquots were
injected onto a X-terra MS C18 column (3.0£150 mm,
5 mm). The mobile phase (20 mM phosphate buffer (pH
´
24. Bort, R.; Mace, K.; Boobis, A.; Gomez-Lechon, M.-J.; Pfeifer,
A.; Castell, J. Biochem. Pharmacol. 1999, 58, 787.

3862
D. Mathieu et al. / Tetrahedron 60 (2004) 3855–3862
25. Hamilton, G. A.; Friedman, J. P. J. Am. Chem. Soc. 1963, 85,
1008.
31. Hamilton, G. J. Am. Chem. Soc. 1964, 86, 3390.
32. Nofre, C.; Cier, A.; Lefier, A. Bull. Soc. Chim. France 1961,
530.
26. Funabiki, T.; Yokomizo, T.; Suzuki, S.; Yoshida, S. Chem.
Commun. 1997, 151.
´
27. Raffard, N.; Balland, V.; Simaan, J.; Letard, S.; Nierlich, M.;
33. Staudinger, Hj.; Ullrich, V. Z. Naturforshg 1964, 19b, 877.
34. Dearden, M. B.; Jefcoate, C. R. E.; Lindsay Smith, J. R. Adv.
Chem. Ser. 1968, 77, 260.
´ `
Miki, K.; Banse, F.; Anxolabehere-Mallart, E.; Girerd, J.-J.
C.R. Chimie 2002, 5, 99.
35. Mizuta, T.; Wang, J.; Miyoshi, K. Inorgan. Chim. Acta 1995,
230, 119.
28. Ullrich, V. Z. Naturforschg 1969, 24b, 699.
29. Mimoun, H.; Seree de Roch, I. Tetrahedron 1975, 31, 777.
30. (a) Funabiki, T.; Toyoda, T.; Ishida, H.; Tsujimoto, M.;
Ozawa, S.; Yoshida, S. J. Mol. Catal. 1990, 61, 235.
(b) Kitajima, N.; Ito, M.; Fukui, H.; Moro-oka, Y. J. Chem.
Soc., Chem. Commun. 1991, 102.
36. Mukerjee, S.; Stassinopoulos, A.; Caradonna, J. P. J. Am.
Chem. Soc. 1997, 119, 8097.