Ferrocenyl catechols: Synthesis, oxidation chemistry and anti-proliferative effects on MDA-MB-231 breast cancer cells

Article abstract of DOI:10.1039/c2dt30700f

The synthesis and anti-tumoral properties of a series of compounds possessing a ferrocenyl group tethered to a catechol via a conjugated system is presented. On MDA-MB-231 breast cancer cell lines, the catechol compounds display a similar or greater anti-proliferative potency (IC₅₀ values ranging from 0.48-1.21 μM) than their corresponding phenolic analogues (0.57-12.7 μM), with the highest activity found for species incorporating the [3]ferrocenophane motif. On the electrochemical timescale, phenolic compounds appear to oxidize to the quinone methide, while catechol moieties form the o-quinone by a similar mechanism. Chemical oxidation of selected compounds with Ag₂O confirms this interpretation and demonstrates the probable involvement of such oxidative metabolites in the in vitro activity of these species.

Full text of DOI:10.1039/c2dt30700f

View Article Online / Journal Homepage / Table of Contents for this issue
Dynamic Article Links
Dalton
Transactions
Cite this: Dalton Trans., 2012, 41, 7537
www.rsc.org/dalton
PAPER
Ferrocenyl catechols: synthesis, oxidation chemistry and anti-proliferative
effects on MDA-MB-231 breast cancer cells†
Yong Leng Kelvin Tan,a,b,c Pascal Pigeon,a,b Siden Top,a,b Eric Labbé,d Olivier Buriez,d Elizabeth A. Hillard,*a,b
Anne Vessières,a,b Christian Amatore,d Weng Kee Leonge and Gérard Jaouen*a,b
Received 28th March 2012, Accepted 26th April 2012
DOI: 10.1039/c2dt30700f
The synthesis and anti-tumoral properties of a series of compounds possessing a ferrocenyl group tethered
to a catechol via a conjugated system is presented. On MDA-MB-231 breast cancer cell lines, the catechol
compounds display a similar or greater anti-proliferative potency (IC50 values ranging from
0.48–1.21 μM) than their corresponding phenolic analogues (0.57–12.7 μM), with the highest activity
found for species incorporating the [3]ferrocenophane motif. On the electrochemical timescale, phenolic
compounds appear to oxidize to the quinone methide, while catechol moieties form the o-quinone by a
similar mechanism. Chemical oxidation of selected compounds with Ag2O confirms this interpretation
and demonstrates the probable involvement of such oxidative metabolites in the in vitro activity of these
species.
enzymes and metal ions to give an o-quinone (OQ) which can
1. Introduction
cause oxidative damage to cellular DNA through redox cycling
with reactive oxygen species such as semi-quinone radicals.10
The selective estrogen receptor modulator (SERM) tamoxifen,
(Z)-2-[4-(1,2-diphenyl-1-butenyl)phenoxy]-N,N-dimethylethana-
mine, is the most commonly prescribed drug for treatment of
hormone-dependent breast cancer.1–3 In addition to its use as a
chemotherapeutic agent, tamoxifen has also been utilized as a
chemopreventive agent for women with a high risk of contract-
ing breast cancer.4 The anti-proliferative action of its metabolite
hydroxytamoxifen on estrogen receptor positive (ER+) cancers
arises due to an anti-estrogenic effect caused by competitive
binding to the ER, resulting in the repression of estradiol-
mediated DNA transcription.5,6 However, tamoxifen suffers from
limitations as it is ineffective against tumors that do not express
the estrogen receptor (ER–). Moreover, the use of tamoxifen has
been associated with an increase in the risk of uterine and endo-
metrial cancer.7,8 Tamoxifen and hydroxytamoxifen can be
further metabolized to the catechol 3,4-dihydroxytamoxifen.9
The latter catechol undergoes oxidation by a variety of oxidative
The OQ is also electrophilic in nature and can cause cell damage
via alkylation of amino acid residues on proteins and covalent
binding to DNA.11
A major research focus of our group is the investigation of the
biological activity of organometallic moieties tethered to
SERMS. We have previously synthesized a series of compounds
which we have termed “hydroxyferrocifens”, by covalently graft-
ing ferrocene onto the hydroxytamoxifen backbone. By modifi-
cation of several structural aspects, these hydroxyferrocifens
have been further developed into the ferrocene phenols, 1a and
2a (Chart 1). Some of these organometallic biovectors demon-
strate a high in vitro anti-proliferative activity, via the combi-
nation of anti-estrogenicity and cytotoxicity on hormone-
dependent (MCF-7) breast cancer cell lines as well as a cytotoxic
effect on hormone-independent (MDA-MB-231) breast cancer
cells.12–15 The strong in vitro activity correlates with the pres-
ence of a ferrocene group, a p-phenol, and a π-system linking
these two functionalities.14,16–19 We have posited that the acti-
vation pathway involves the oxidation of the phenol moiety via
the ferrocene group to generate an active quinone methide (QM)
species, and have shown that these QMs were indeed formed in
the presence of rat liver microsomes as well as during chemical
oxidation with silver oxide.20,21 More recently, we have syn-
thesized cyclic ferrocenyl compounds based on the [3]ferroceno-
phane motif, 1b and 2b, and found them to be considerably
more potent than their non-cyclic analogues 1a and 2a.22–24 As
part of our ongoing investigations into the biological activity of
the above-mentioned compounds, we were interested in examin-
ing the in vitro effects of their catechol analogues. In this article,
we present the synthesis, electrochemical and chemical oxidation
aENSCP Chimie Paris Tech, Laboratoire Charles Friedel (LCF), 75005
Paris, France
bCNRS, UMR 7223, 75005 Paris, France. E-mail:
gerard-jaouen@enscp.fr; Tel: +33 (0)1 43 26 95 55; E-mail:
hillard@crpp-bordeaux.cnrs.fr; Tel: +33 (0)5 56 84 56 24
cHwa Chong Institution, 661 Bukit Timah Road, Singapore 269734,
Singapore
dEcole Normale Supérieure, Département de Chimie, 24 rue Lhomond,
75231 Paris cedex 05, France
eDivision of Chemistry and Biological Chemistry, School of Physical
and Mathematical Sciences, Nanyang Technological University,
21 Nanyang Link, Singapore 637371, Singapore
†Electronic supplementary information (ESI) available: CCDC
829115–829118. For ESI and crystallographic data in CIF or other elec-
tronic format see DOI: 10.1039/c2dt30700f
This journal is © The Royal Society of Chemistry 2012
Dalton Trans., 2012, 41, 7537–7549 | 7537

View Article Online
studies for a series of catechol ferrocene and [3]ferrocenophane
compounds. The results of biological tests to evaluate the anti-
proliferative effects of these catechol compounds on hormone-
independent breast cancer cell lines, as compared to their pheno-
lic analogues, will also be reported.
Results and discussion
2.1 Synthesis and characterization
Previous studies have shown that the diphenols 1a and 1b are
more cytotoxic than their monophenol analogues 2a and 2b. We
were curious to know if the presence of two hydroxyl groups on
the same phenyl ring would similarly enhance the activity of the
ferrocene tamoxifen derivatives. To approach this question, we
naturally began by synthesizing the catechol analogues of 2a and
2b (compounds 3a and 3b, Chart 2). The phenol (2c) and cate-
chol (3c) compounds where a phenyl group replaces the CH3
group were also prepared with the assumption that the phenyl
substituent could favor QM formation by extending the conju-
gated system and thus stabilizing the oxidation product. As com-
pounds 3a, 3b and 3c could theoretically be metabolized to
either a QM or an OQ, we also prepared the phenol compound
2d and the catechol 3d where the lack of a labile proton β to the
ferrocene group would prevent extended QM formation.
A McMurry cross-coupling between 4-hydroxybenzophenone
or 3,4-dihydroxybenzophenone with [3]ferrocenophan-1-one or
the appropriate ferrocenyl ketone provided the general synthetic
route to all of the ferrocenyl phenol and catechol compounds
shown in Chart 2, which were obtained in moderate to good
yields. For example, a Friedel–Crafts reaction of propionyl chlor-
ide with ferrocene afforded ethyl ferrocenyl ketone, which was
further combined with 3,4-hydroxybenzophenone to afford the
dark red compound 3a in 65% yield (Scheme 1).25 Efforts to
obtain the compound with vicinal phenol groups (2f) were
unsuccessful. The mixture obtained from the McMurry coupling
reaction between 4-hydroxyphenyl ferrocenyl ketone and
4-hydroxybenzophenone was found to decompose rapidly
during the process of chromatographic separation on a silica gel
column, and attempts to deprotect the phenol in 2e by removing
the methyl group using boron tribromide at ambient temperature
Chart 1 The SERM hydroxytamoxifen and its ferrocenyl derivatives.
Chart 2 New ferrocenyl tamoxifen derivatives studied in this report.
7538 | Dalton Trans., 2012, 41, 7537–7549
This journal is © The Royal Society of Chemistry 2012

View Article Online
Scheme 1 (a) Synthesis of 3a by McMurry cross coupling (similar procedure used for synthesis of 2c–e and 3a–d) and; (b) Neither McMurry cross
coupling nor deacetylation yielded the desired compound 2f; (c) Acetylation of compound 3a yielded 4, 5 and 6.
also resulted in decomposition of the reaction mixture. An
endeavor to obtain the catechol analogue of 2f, viz. 3f, by react-
ing 4-hydroxyphenyl ferrocenyl ketone and 3,4-dihydroxybenzo-
phenone also failed to give the expected product. The reaction of
3a with an excess of acetyl chloride at ambient temperature
afforded the diacetoxy derivative 4, as well as the monoacetoxy
derivatives 5 and 6; the latter two isomeric compounds could not
be separated by chromatography or on semi-preparative HPLC.
All the novel compounds were characterized spectroscopically
and analytically. Molecular ion peaks are present in all their
respective mass spectra; for the acetylated derivatives 4–6, peaks
corresponding to loss of CH3CO˙ fragments are displayed as
well. Consistent with previous observations, the McMurry reac-
tion resulted in both the Z- and E-isomers of the tetrasubstituted
olefins. We were unable to separate the geometric isomers of
each compound by chromatography, and the presence of both
isomers was observed in the solution 1H NMR spectra. After
fractional crystallization, compounds E-2c, Z-2d, Z-2e and Z-4
were studied by single-crystal X-ray diffraction methods, and the
ORTEP diagrams of E-2c, Z-2e and Z-4 are illustrated in Fig. 1
(Z-2d omitted for brevity). A common atomic numbering
scheme, together with selected bond parameters for all four com-
pounds are collected in Table 1. For each molecule, there is a
narrowing of the bond angle between the two substituents
connected to the same carbon atom of the double bond, presum-
ably to minimize steric interaction with their cis-disposed neigh-
bours. While the planarity of the ethylenic skeleton is mainly
retained, the phenyl moieties in each molecule are tilted with
respect to the olefinic plane, ranging from 44.5(2)° in Z-2e
to 71.5(2)° in Z-2d. The cyclopentadienyl ring attached to
the double bond also displays deviation from the latter plane,
with observed twist angles between 13.6(2)° in Z-2d to 31.5(2)°
in Z-4.
2.2 Electrochemistry
We previously hypothesized that the anti-proliferative effects of
several ferrocenyl phenols, such as 1a and 2a, are associated
with their in situ oxidation to QM species.20,21,26 The ferrocene
moiety seems to act as an intramolecular oxidant via a proton-
coupled electron transfer process which can be observed on the
electrochemical timescale. In order for QM formation to occur
via this process, the molecules must possess an appropriate geo-
metry; namely the presence of a ferrocene group conjugated to a
p-phenol, with a labile proton β to the ferrocene group. For
example, 2c differs from 2a only by the replacement of the ethyl
group by a benzyl moiety. Therefore, all of the requirements for
This journal is © The Royal Society of Chemistry 2012
Dalton Trans., 2012, 41, 7537–7549 | 7539

View Article Online
absence of added 2,6-dimethylpyridine is similar to what was
previously observed for 2a in methanol and pyridine (ESI†).‡
On the other hand, compounds 2d, 2e and 3d lack such a labile
proton. While neither 2d nor 2e show reactivity with the added
base, 3d displays an electrochemical signature suggestive of
proton-coupled electron transfer. While a QM structure is not
apparently feasible for 3d, it should be noted that the o-quinone
is also a possible two-electron oxidation product. Catechol com-
pounds 3a–c, possessing a labile proton β to the ferrocene
group, can theoretically form either the QM or the o-quinone
after a two-electron oxidation. The cyclic voltammograms for
3a–c more closely resemble that of 3d than 2a, suggesting that
the o-quinone may be formed on the electrochemical timescale
(ESI†).
The electrochemical behavior of 3a and 3c was further
explored by cyclic voltammetry in acetonitrile.† The voltammo-
grams obtained in the absence and presence of imidazole are pre-
sented in Fig. 2.
In the absence of imidazole, one can observe the one-electron
oxidation of 3a at a peak potential value of 0.42 V/SCE. This
electrochemical step is ascribed to the reversible oxidation of the
ferrocene group (Fc) to ferricenium (Fc+), in agreement with the
behavior of analogous ferrocenyl tamoxifen derivatives.20 In the
presence of imidazole, two oxidation steps are observed; a
nearly 2-electron irreversible oxidation peak appears, the poten-
tial of which shifts towards less positive values as the imidazole
concentration increases. Another peak located at 0.52 V/SCE
features a 1-electron reversible oxidation process, its potential
being independent of imidazole concentration. The CVs
obtained with 3c (right panel) under the same conditions are
akin to those for 3a (left panel). We recently carried out a study
devoted to the determination of the oxidation mechanism of a
model ferrocene phenol, which showed a very similar base-
dependent electrochemical behavior.27 Parallel EPR and electro-
chemical monitoring of electrolyses established that the first
2-electron oxidation process corresponds to the passage from a
phenol to a QM, followed by the 1-electron oxidation of the fer-
rocene group of the latter species. The base (imidazole) triggers
an oxidation sequence made possible by the increased acidity of
the phenol proton of the ferricenium cation generated upon oxi-
dation of the starting ferrocene phenol compound. In the present
study, an OQ could be generated instead of a QM, as featured in
Scheme 2.
In the absence of imidazole, one only observes the reversible
one-electron oxidation of 3a and 3c (step (1)). In the presence of
excess imidazole, a 2-electron oxidation sequence occurs accord-
ing to steps (1) + (2) + (3). At low scan rates (Fig. 2), the poten-
tial dependence of this bielectronic wave with excess imidazole
features the thermodynamically favored passage from a mono-
electronic Fc oxidation to a proton-coupled bielectronic oxi-
dation of the catechol. Note that the deprotonations are most
‡Although our earliest studies attempted to model the cellular environ-
ment using methanol (the lipophilic ferrocifens being essentially insolu-
ble in water), we quickly realized that other organic solvents have
certain practical advantages, and are not less appropriate than methanol
when considering that these compounds are probably activated by a base
found not in the cytoplasm but in a biomolecule. It should be mentioned
that no solvent dependence has been observed in the electrochemical
studies of ferrocifen derivatives.
Fig. 1 ORTEP diagrams (50% probability thermal ellipsoids, organic
hydrogen atoms omitted) for E-2c, Z-2e and Z-4.
ferrocene-mediated QM formation are met, and, indeed, the
cyclic voltammetry in wet dichloromethane in the presence and
7540 | Dalton Trans., 2012, 41, 7537–7549
This journal is © The Royal Society of Chemistry 2012

View Article Online
Table 1 Common atomic numbering scheme and selected bond parameters for E-2c, Z-2d, Z-2e and Z-4
E-2c
Z-2d
Z-2e
Z-4
Bond lengths (Å)
C1–C2
C1–C3
C1–C4
C2–C5
1.343(5)
1.528(4)
1.481(4)
1.496(5)
1.498(4)
1.349(4)
1.503(4)
1.482(4)
1.494(4)
1.500(3)
1.360(3)
1.499(3)
1.473(3)
1.490(3)
1.489(3)
1.345(3)
1.526(3)
1.475(3)
1.499(3)
1.495(3)
C2–C6
Bond angles (°)
C3–C1–C4
C5–C2–C6
C3–C1–C2–C5
C4–C1–C2–C6
C3–C1–C2–C6
114.6(3)
113.0(3)
4.3(3)
6.1(3)
−170.3(2)
115.3(2)
114.1(2)
10.0(2)
2.5(2)
−171.4(2)
114.95(18)
114.66(17)
19.5(2)
11.4(2)
−159.5(2)
115.35(16)
112.98(16)
2.5(2)
0.1(2)
−173.3(2)
Fig. 2 Cyclic voltammograms of 3a (left) and 3c (right) 1 mM in acetonitrile + Bu4NBF4 (100 mM). Platinum electrode (∅ = 0.5 mm), scan rate
0.2 V s−1. CVs recorded in the absence (black solid line) and in the presence of 1 molar equivalent (black dashed line) and 10 molar equivalents
(grey solid line) of imidazole.
likely concerted to the electron transfers, at least within the time
range explored. The following one-electron process corresponds
to the reversible oxidation of the orthoquinone into its ferrice-
nium analogue (step (4)). Moreover, the (1) + (2) + (3) sequence
is fast enough to be observed within the same scan in the pres-
ence of imidazole.
In order to confirm this oxidation mechanism, we performed
preparative electrolyses of catechols 3a and 3c in the presence of
imidazole at 0.48 V/SCE, i.e. the potential value corresponding
to the diffusion tail of the first bielectronic wave and to the foot
of the second monoelectronic wave. The concentration of the
species was monitored by cyclic voltammetry, as reported in Fig. 3.
For both compounds, the complete electrolysis at 0.48 V/SCE
in the presence of imidazole corresponds to a total charge of ca.
2 Faraday mol−1. Curves c and c′, recorded after the electrolyses,
no longer display the original 2-electron oxidation wave (as seen
in curves b and b′ recorded before the electrolyses) and the CVs
only show the reversible one-electron oxidation ascribed to the
OQ. Finally, the bielectronic nature of the first wave is quantitat-
ively confirmed at the time scale of preparative electrolyses.
The electrochemical oxidation of catechols 3a and 3c is dra-
matically affected by the presence of imidazole. In the absence
of this base, both catechols only oxidize according to a reversible
one-electron process centered on the ferrocene group. In the
presence of imidazole, the CVs display two distinct processes,
the first one corresponding to the bielectronic oxidation of the
This journal is © The Royal Society of Chemistry 2012
Dalton Trans., 2012, 41, 7537–7549 | 7541

View Article Online
Scheme 2 Base-promoted oxidation sequence of catechols 3a and 3c.
Fig. 3 Cyclic voltammograms of acetonitrile + Bu4NBF4 (200 mM) solutions (platinum electrode ∅ = 0.5 mm, scan rate 0.1 V s−1) at various elec-
trolysis stages. Left panel: Compound 3a 1.2 mM (curve a); Compound 3a 1.2 mM + imidazole 15 mM (curve b); and Compound 3a 1.2 mM + imi-
dazole 15 mM + 1.95 Faraday mol−1 electrolysis at 0.48 V/SCE (curve c) Right panel: Compound 3c 6 mM (curve a′); Compound 3c 6 mM +
imidazole 75mM (curve b′); and Compound 3c 6 mM + imidazole 75 mM + 1.8 Faraday mol−1 electrolysis at 0.48 V/SCE (curve c′). Insets = evol-
ution of the charge vs. time during electrolysis.
catechol to the OQ, the second one being the ferrocene-centered
one-electron reversible oxidation of the OQ formed at the first
wave. The oxidation sequence, although leading to OQs, dis-
plays kinetic and thermodynamic features very similar to those
encountered in the electrochemical oxidation of ferrocene
phenols to QMs.
the previously reported transformation of compounds such as 1a
to the quinone methide using Ag2O probably does not involve
the intramolecular electron transfer process observed in the elec-
trochemical experiments. Nonetheless, the use of Ag2O allows
us to determine whether or not a QM or OQ can ultimately be
formed and to isolate such oxidation products.
In contrast to the unreactive 2d and 2e, chemical oxidation of
2c resulted in the formation of the QM 7 (Scheme 3), in accord
with the electrochemical observations. Compound 7 was ident-
ified by the presence of a carbonyl stretch at 1709 cm−1 in the
infrared spectrum, as well as a molecular ion peak (m/z 468)
indicating the loss of two hydrogen atoms from the parent com-
pound 2c (m/z 470). For the 13C NMR spectrum of 7, the reson-
ance at δ 187.2 ppm is ascribable to the presence of a CvO
2.3 Chemical oxidation
The reaction of 2d and 2e, lacking a labile proton β to the ferro-
cenyl group, with silver oxide in acetone afforded only unreacted
starting compounds, as expected. Interestingly, Ag2O does not
directly oxidize the ferrocene group to ferricenium, and therefore
7542 | Dalton Trans., 2012, 41, 7537–7549
This journal is © The Royal Society of Chemistry 2012

View Article Online
Scheme 3 Chemical oxidation of compounds 2c and 3a–d.
1
group. In the H NMR spectrum, the multiplet signals in the δ
cyclopentadienyl rings of the Z-isomer than that of the E-isomer,
analysis of the chemical shifts in the proton NMR spectra
suggests that the major isomers of 8a, 8b and 8c have the Z
configuration while 8d has the E configuration.
6.24–7.25 ppm range, which integrate to a total of five hydrogen
atoms, are assigned to the protons of the quinone ring as well as
the single vinylic proton, while the more deshielded multiplets
in the δ 7.41–7.67 ppm range, which integrate to a total of ten
protons, can be attributed to the phenyl rings.
The catechol compounds 3a–d reacted with silver oxide in
acetone to give the corresponding o-quinones 8a–d, as suggested
by the electrochemical experiments. While quantitative for-
mation of 8b–d was observed after twenty minutes, the complete
conversion of 3a to 8a took one hour. At twenty minutes, the
oxidation of 3a yielded a mixture of 8a, the QM as well as start-
ing material; the ratio of 8a to the latter two compounds was
approximately 9 : 1. Carbonyl stretches at 1650 cm−1 are dis-
played in the infrared spectra of the OQs. Singlet signals are also
present in the δ 180 ppm region of the 13C NMR spectra. The
1H NMR spectra of 8a–d indicate that the protons of the
o-quinone ring are shielded as compared to the protons of the cate-
chol ring in 3a–d. In addition, the proton signals attributable to
the cyclopentadienyl rings of the o-quinone compounds are
deshielded relative to their catechol precursors (Table 2). Assum-
ing that the proximity of the o-quinone ring would result in a
greater impact on the chemical shift of the protons of the
2.4 Effect on growth of hormone-independent breast cancer
cells MDA-MB-231
Compounds 2, 3 and 4 were screened for their anti-proliferative
effects on the MDA-MB-231 breast cancer cell line, and the
results are recorded in Table 3. This line is hormone independent
and does not express the estrogen receptor, hence on these cell
lines the observed anti-proliferative effect can be attributed only
to a cytotoxic effect potentially induced by the ferrocenyl unit.
Compounds 2d and 2e, which remained inert to chemical oxi-
dation and cannot undergo ferrocene-mediated QM formation,
were markedly less active on the cells as compared to 2a–c
which were able to be oxidized to quinone species; this provides
further support for our proposal that oxidation to QM moieties
could be the key to the anti-proliferative effect displayed by
these ferrocenyl compounds. Along these same lines, the lower
activity of 2c can be attributed to the extended conjugated
This journal is © The Royal Society of Chemistry 2012
Dalton Trans., 2012, 41, 7537–7549 | 7543

View Article Online
Table 2 NMR assignments for 3a–d and 8a–d (cyclopentadienyl
rings, in CD3COCD3)
active. The steric bulk of two acetyl groups ortho to one another
may prevent the protecting groups from being hydrolyzed in the
cell.
δH (ppm)
Compound C5H4
δH (ppm)
Compound C5H4
The ferrocenyl catechols have similar (3a and 3b) or stronger
(3c and 3d) anti-proliferative potency to their corresponding phe-
nolic analogues 2a–d. This is in contrast to previous reports that
3,4-dihydroxytamoxifen and 3,4-dihydroxytoremifen had lower
biological activity than 4-hydroxytamoxifen and 4-hydroxito-
remifen, respectively, leading the authors to conclude that the
metabolism of tamoxifen and toremifen to their catechols might
play a minor role in the cytotoxicity of the purely organic mol-
ecules.10,30 The addition of ferrocene thus reverses the trend
such that the catechols show similar activity towards the cancer
cells as the phenolic compounds. Compound 3b, with the [3]fer-
rocenophane structure, displays the highest anti-proliferative
effect amongst the catechol complexes. This is consistent with
our previous observations on these types of molecules, and the
enhanced efficacy of the compounds bearing the [3]ferroceno-
phane motif may be attributed to the narrower HOMO–LUMO
gap present in the oxidized form of the [3]ferrocenophane moi-
eties, giving rise to more reactive species.31 Compounds 2d, 2e
and 4 cannot form either the QM or the OQ via ferrocene-
mediated oxidation due to their geometry. These compounds
have the lowest activity of the series, suggesting that ferrocene is
an important mediator of the oxidation process. We have postu-
lated that ferrocenyl group acts as a redox antenna that triggers
the oxidation of the phenol function in conjugated systems. The
present results support this hypothesis, and we furthermore see
that the ferrocenyl group also plays an important role in the oxi-
dation of the catechol derivatives. The molecule can transform
into a quinone methide or an orthoquinone, depending on the
location of the most acidic proton, and the ease of the phenol
oxidation by delocalization of the radical is the key to the anti-
proliferative activity of compounds.
C5H5
C5H5
(Z+E)-3a
4.04 (2H), 4.11
3.84 (2H)a (5H)a
4.07 (2H), 4.12
Z-8ab
4.58 (2H), 4.28
4.53 (2H)a (5H)a
3.96 (2H)
(5H)
E-8ab
Z-8bb
4.18 (2H), 4.17
3.91 (2H)
4.49 (2H),
4.46 (2H)a
4.25 (2H),
3.98 (2H)a
4.37 (2H),
4.07 (2H),
4.04 (2H),
3.94 (2H)
(5H)
—
(Z+E)-3b
4.26 (2H),
4.04 (2H)a
4.00 (2H),
3.94 (2H)a
4.24 (2H),
4.02 (2H),
3.97 (2H),
3.96 (2H)
E-8bb
Z-8cb
—
(Z+E)-3c
(Z+E)-3d
3.99 (2H), 4.13
3.82 (2H)a (5H)a
3.94 (2H), 4.16
4.45 (4H)a 4.35
(5H)a
3.69 (2H)
(5H)
E-8cb
Z-8db
4.15 (2H), 4.24
3.88 (2H)
4.48 (2H), 4.28
4.30 (2H) (5H)
(5H)
4.06 (2H), 4.15
3.54 (2H)a (5H)a
4.01 (2H), 4.12
3.39 (2H)
(5H)
E-8db
4.24 (2H), 4.25
3.51 (2H)a (5H)a
a Peaks belonging to major isomer. b Assignments of isomers performed
based on δ (see text).
Table 3 Cytotoxicity of 1a, 2a–e, 3a–d and 4 on MDA-MB-231
breast cancer cell lines
Compound Phenol or catechol IC50 (μM)a Presumed metaboliteb
(Z+E)-1a
(Z+E)-2a
(Z+E)-2b
(Z+E)-2c
(Z+E)-2d
(Z+E)-2e
(Z+E)-3a
(Z+E)-3b
(Z+E)-3c
(Z+E)-3d
(Z+E)-4
2 phenols
1 phenol
1 phenol
1 phenol
1 phenol
1 phenol
Catechol
Catechol
Catechol
Catechol
0.29 0.07 QM
1.80 0.03 QM
0.57 0.03 QM
Experimental
3.1 General procedures
5.3 0.9
9.0 1.0
12.7 1.6
QM
None
None
All reactions and manipulations were carried out under an argon
atmosphere using standard Schlenk techniques. THF was dis-
tilled over sodium–benzophenone prior to use. Column chrom-
atography was performed on silica gel 60 M. Semi-preparative
HPLC separations were performed on a Shimadzu instrument
with a Kromasil C18 column (length of 25 cm, diameter of
2 cm, particle sizes of 10 μm) using acetonitrile as eluent. IR
spectra were obtained on a BOMEM Michelson 100 spec-
1.10 0.07 OQ
0.48 0.04 OQ
1.21 0.07 OQ
2.4 0.2
OQ
None
Protected catechol 15.1 1.5
a Mean of two experiments range. b As determined by electrochemical
or chemical oxidation.
1
trometer. H and 13C NMR spectra were acquired on a Bruker
300 MHz spectrometer. The solvent used was deuterated chloro-
form unless otherwise stated. Mass spectrometry was carried out
at the Service de Spectrométrie de Masse at ENSCP, Paris. High
resolution mass spectra (HRMS) were obtained by the Groupe
de Spectroscopie de Masse of the Laboratoire de Structure et
Fonction de Molécules Bioactives at the University of Pierre et
Marie Curie, Paris. Microanalyses and biochemical experiments
were performed, respectively, by the Service de Microanalyse
and IMAGIF at the ICSN (Gif sur Yvette, France).
system that would be expected to stabilize the QM towards
nucleophilic attack in the cell. It should be mentioned that the
IC50 values of 1a (0.6 μM)28 and 2a (1.13 μM)22 have been pre-
viously determined, using a 24-well plate after five days. The
currently reported values, obtained using a 96-well plate after
72 h vary somewhat, but are on the same order of magnitude.
Interestingly, although previous work showed that 1a and its
acetylated form show equal activity against MDA-MB-231
cells,29 compound 4, the acetylated derivative of 3a, is not
Benzyl ferrocenyl ketone, phenyl ferrocenyl ketone, 4-meth-
oxyphenyl ferrocenyl ketone, ethyl ferrocenyl ketone and
7544 | Dalton Trans., 2012, 41, 7537–7549
This journal is © The Royal Society of Chemistry 2012

View Article Online
[3]ferrocenophan-1-one were synthesized according to pro-
cedures found in the literature.32–34 All other reagents were from
commercial sources and used as supplied.
semi-preparative HPLC with acetonitrile as the eluent. The
acetonitrile was removed under reduced pressure and the product
was recrystallized from a hexane–CH2Cl2 solution.
3.4 Oxidation reactions
3.2 Reactions of ferrocenyl ketones with benzophenones
3.4.1 Oxidation of 2c. Silver oxide (0.10 g, 0.4 mmol) was
added to a solution of 2c (0.10 g, 0.2 mmol) in 8 mL of acetone
and the mixture was stirred at room temperature in the dark. At
hourly intervals a sample was withdrawn, filtered and the solvent
removed under reduced pressure, after which NMR spectro-
scopic analysis was carried out. Quantitative formation of 7 was
observed after 2 h. 7: IR(CH2Cl2) ν(CO) (cm−1): 1709. Other
spectroscopic and analytical data are given in Tables 5 and 6.
In a typical reaction, TiCl4 was added dropwise to a suspension
of zinc powder in 80 mL of THF and the mixture was heated at
reflux for 2 h. A second solution was prepared by dissolving the
ferrocenyl ketone and benzophenone in THF (30 mL). The latter
solution was added dropwise to the first solution and the result-
ing mixture was heated at reflux. After cooling to room tempera-
ture, the mixture was stirred with water and CH2Cl2. The
mixture was acidified with dilute HCl until the dark color dis-
appeared and was decanted. The aqueous layer was extracted with
CH2Cl2 and the combination of organic layers dried on MgSO4.
This was followed by filtering the mixture and removing the
solvent under reduced pressure. Chromatographic separation on
a silica gel column with hexane–CH2Cl2 as the eluent afforded
compounds 2c–e and 3a–d respectively. The reaction conditions
and yields are summarized in Table 4, and spectroscopic and
analytical data for the compounds are given in Tables 5 and 6.
For the biological tests, each compound was re-purified on semi-
preparative HPLC with acetonitrile as the eluent. The acetonitrile
was removed under reduced pressure and the product was recrys-
tallized from a hexane–CH2Cl2 solution.
3.4.2 Oxidation of 3a–d. In a typical reaction, silver oxide
(0.12 g, 0.5 mmol) was added to a solution of the ferrocenyl
catechol (0.1 mmol) in 3 mL (CD3)2CO and the mixture was
stirred at room temperature. Quantitative formation of 8b–d was
observed after 20 min; 8a after 1 h. IR(CH2Cl2) ν(CO) (cm−1):
1647 (8a); 1658 (8b); 1651 (8c); 1650 (8d). NMR data are given
in Table 5.
3.5 Electrochemistry
Cyclic voltammograms (CVs) were obtained using a three elec-
trode cell with a 0.5 mm Pt working electrode, stainless steel rod
counter electrode, and Ag/AgCl ethanol reference electrode, with
an μ-Autolab 3 potentiostat driven by GPES software (General
Purpose Electrochemical System, Version 4.8, EcoChemie B.V.,
Utrecht, the Netherlands). Solutions consisted of 10 mL CH2Cl2,
approximately 1 mM analyte, and 0.1 M Bu4NPF6 supporting
electrolyte. After obtaining the cyclic voltammograms in
CH2Cl2, two drops of lutidine were added to the cell and data
was obtained.
Preparative electrolyses were carried out in acetonitrile under
argon atmosphere in a 2 × 5 cm3 two-compartment cell with a n°
4 porosity sintered glass separator. The anodic compartment was
equipped with a platinum grid anode (1 cm2) and a SCE refer-
ence electrode, the cathodic compartment being fitted with a
1 cm2 area platinum grid cathode. The compartments were filled
3.3 Acetylation of 3a
NaH (0.34 g, 8.50 mmol, 60% in oil) was added to a solution of
3a (1.50 g, 3.53 mmol) dissolved in dry THF (60 mL). Acetyl
chloride (0.55 mL, 7.80 mmol) was added and the mixture was
stirred overnight at ambient temperature, after which 2 mL of
ethanol was added. After stirring for 10 min the solution was
poured into water and extracted with CH2Cl2. The combined
organic layers were dried on MgSO4, filtered and the solvent
was removed under reduced pressure. Chromatographic separ-
ation on a silica gel column with hexane–CH2Cl2 as the eluent
afforded 4 (0.64 g, 36%) as well as a mixture of 5 and 6 (0.51 g,
31%), respectively. For the biological tests, 4 was re-purified on
Table 4 Amount of reagents used, products and yields
Reaction
time
Reaction Zinc
TiCl4
Ferrocenyl ketone
Benzophenone
Product Yield
1
2
3
4
5
6
7
1.96 g
(30 mmol)
1.96 g
(30 mmol)
1.96 g
(30 mmol)
7.84 g
(120 mmol)
1.96 g
(30 mmol)
3.92 g
(60 mmol)
2.94 g
(45 mmol)
3.79 g, 2.2 mL
(20 mmol)
3.79 g, 2.2 mL
(20 mmol)
3.79 g, 2.2 mL
(20 mmol)
15.16 g, 8.8 mL Ethyl ferrocenyl ketone
(80 mmol)
3.79 g, 2.2 mL
(20 mmol)
7.58 g, 4.4 mL
(40 mmol)
11.37 g, 3.3 mL Phenyl ferrocenyl ketone
(30 mmol) 2.17 g (7.5 mmol)
Benzyl ferrocenyl ketone
1.52 g (5 mmol)
Phenyl ferrocenyl ketone
1.45 g (5 mmol)
4-Methoxyphenyl ferrocenyl 4-Hydroxybenzophenone 0.99 g
ketone 1.60 g (5 mmol)
4-Hydroxybenzophenone 0.99 g
(5 mmol)
4-Hydroxybenzophenone 0.99 g
(5 mmol)
Overnight
4 h
2c
2d
2e
3a
3b
3c
3d
1.04 g
(44%)
1.19 g
(52%)
0.88 g
(36%)
5.52 g
(65%)
0.99 g
(47%)
1.29 g
(27%)
1.47 g
(42%)
Overnight
2 h
(5 mmol)
3,4-Dihydroxybenzophenone
4.28 g (20 mmol)
3,4-Dihydroxybenzophenone
1.07 g (5 mmol)
3,4-Dihydroxybenzophenone
2.14 g (10 mmol)
4.84 g (20 mmol)
[3]Ferrocenophan-1-one
1.20 g (5 mmol)
Benzyl ferrocenyl ketone
3.04 g (10 mmol)
4 h
Overnight
4 h
3,4-Dihydroxybenzophenone
1.61 g (7.5 mmol)
This journal is © The Royal Society of Chemistry 2012
Dalton Trans., 2012, 41, 7537–7549 | 7545

View Article Online
Table 5 NMR data for compounds 2c–e and 3–8
Compound 1H NMR, δ (ppm)
13C NMR, δ/ppm
2c
2d
2e
7.16–7.08 (m, 10H, 2C6H5), 6.96 (d, J = 7.9 Hz, 2H, C6H4), 6.64
153.3, 153.2, 143.7, 143.4, 140.9, 139.4, 136.1, 135.9, 132.5,
(d, J = 7.9 Hz, 2H, C6H4), 4.68 and 4.64 (s, br, 1H, OH), 4.12 and 132.1 (Cquat), 130.6, 130.0, 129.2, 128.7, 127.5, 127.4, 127.3,
4.10 (s, 5H, C5H5), 3.97 and 3.94 (t, J = 1.9 Hz, 2H, C5H4), 3.91
and 3.87 (s, 2H, CH2), 3.80 and 3.74 (t, J = 1.9 Hz, 2H, C5H4)
[isomers present in 1 : 1 ratio]
7.37–7.07 (m, 10H, 2C6H5), 6.90 and 6.82 (d, J = 7.3 Hz, 2H,
C6H4), 6.69 and 6.37 (d, J = 7.3 Hz, 2H, C6H4), 4.66 (s, br, 1H,
OH), 4.35 and 4.33 (t, J = 1.9 Hz, 2H, C5H4), 4.30 (s, 5H, C5H5),
4.19 and 4.13 (t, J = 1.9 Hz, 2H, C5H4) [isomers present in 1 : 1
ratio]
126.9, 125.6, 125.5, 124.8, 114.3 (C6H5, C6H4), 87.5, 87.2
(Fcipso), 69.4, 69.3 (C5H4), 68.8, 68.7 (C5H5), 67.6, 67.5 (C5H4),
40.0, 39.9 (CH2)
145.2, 145.0, 144.3, 144.1, 140.0, 139.3, 138.5, 138.4 (Cquat),
133.4, 133.1, 132.0, 131.7, 130.5, 129.3, 129.2, 128.8, 128.2,
127.5, 117.5, 116.2 (C6H5, C6H4), 86.8, 85.9 (Fcipso), 72.2,
72.0 (C5H4), 71.9, 71.8 (C5H5), 71.0, 70.9 (C5H4)
7.35–7.09 (m, 5H, C6H5), 6.90 (d, J = 8.9 Hz, 2H, C6H4), 6.87 (d, 157.9, 154.4, 153.1, 144.5, 138.7, 138.6, 136.9, 134.8, 134.7
J = 8.9 Hz, 2H, C6H4), 6.79 (d, J = 8.5 Hz, 2H, C6H4), 6.40 (d, J
= 8.5 Hz, 2H, C6H4), 4.40 (s, br, 1H, OH), 4.07 (s, 5H, C5H5),
4.01 (t, J = 1.9 Hz, 2H, C5H4), 3.98 (t, J = 1.9 Hz, 2H, C5H4),
3.70 (s, 3H, OCH3) [major isomer]
(Cquat), 132.2, 131.4, 131.1, 130.1, 129.8, 128.6, 127.3, 126.7,
125.4, 115.5, 114.3, 112.8, 112.7 (C6H5, C6H4), 86.4 (Fcipso),
70.0, 69.9 (C5H4), 69.3 (C5H5), 68.4, 68.3 (C5H4), 55.1 (OCH3)
7.35–7.09 (m, 5H, C6H5), 6.93 (d, J = 8.7 Hz, 2H, C6H4), 6.75 (d,
J = 8.7 Hz, 2H, C6H4), 6.70 (d, J = 8.9 Hz, 2H, C6H4), 6.68 (d, J
= 8.9 Hz, 2H, C6H4), 4.68 (s, br, 1H, OH), 4.08 (s, 5H, C5H5),
3.73 (s, 3H, OCH3), 3.51 (t, J = 1.9 Hz, 2H, C5H4), 3.43
(t, J = 1.9 Hz, 2H, C5H4) [minor isomer]
3aa
7.35–7.04 (m, 5H, C6H5), 6.84 (d, J = 8.2 Hz, 1H, C6H3), 6.70 (d, 144.8, 144.6, 143.1, 143.3, 142.1, 142.0 (Cquat), 129.8, 129.3,
J = 2.0, 1H, C6H3), 6.60 (dd, J = 8.2 and 2.0 Hz, 1H, C6H3), 4.11
(s, 5H, C5H5), 4.04 (t, J = 1.9 Hz, 2H, C5H4), 3.84 (t, J = 1.9 Hz,
128.2, 128.1, 126.1, 122.8, 122.6, 122.1 (C6H5, C6H3), 86.6, 85.7
(Fcipso), 69.3 (C5H4), 69.2 (C5H5), 68.1 (C5H4), 27.9 (CH2), 15.6,
2H, C5H4), 2.65 (q, J = 7.5 Hz, 2H, CH2), 1.05 (t, J = 7.5 Hz, 3H, 15.5 (CH3)
CH3) [major isomer]
7.35–7.04 (m, 5H, C6H5), 6.71 (d, J = 8.0 Hz, 1H, C6H3), 6.57 (d,
J = 2.0, 1H, C6H3), 6.43 (dd, J = 8.0 and 2.0 Hz, 1H, C6H3), 4.12
(s, 5H, C5H5), 4.07 (t, J = 1.9 Hz, 2H, C5H4), 3.96 (t, J = 1.9 Hz,
2H, C5H4), 2.55 (q, J = 7.5 Hz, 2H, CH2), 1.03 (t, J = 7.5 Hz, 3H,
CH3) [minor isomer]
3a partially decomposed within 6 months. After purification by
chromatography, the ratio between the two isomers evolved from
1 : 1 to 78 : 12.
3ba
7.63 (s, 1H, OH), 7.57 (s, 1H, OH), 7.38–7.22 (m, 5H, C6H5),
143.8, 142.2, 142.1, 136.5, 134.1, 133.9 (Cquat), 129.4, 128.3,
6.53 (d, 3J = 8.1 Hz, 1H, C6H3), 6.51 (d, 4J = 2.1 Hz, 1H, C6H3),
126.8, 124.0, 117.9, 114.5 (C6H5, C6H3), 87.1 (Fcipso), 70.6, 70.5,
6.39 (dd, 3J = 8.1 Hz, 4J = 2.1 Hz, 1H, C6H3), 4.26 (s, 2H, C5H4), 69.4, 68.9 (C5H4), 40.9, 28.8 (CH2)
4.04 (s, 2H, C5H4), 4.00 (s, 2H, C5H4), 3.94 (s, 2H, C5H4), 2.61
(m, 2H, CH2), 2.30 (m, 2H, CH2) [major isomer]
7.86 (s, 1H, OH), 7.84 (s, 1H, OH), 7.06–7.01 (m, 5H, C6H5),
6.82 (d, 3J = 8.1 Hz, 1H, C6H3), 6.70 (d, 4J = 2.1 Hz, 1H, C6H3),
6.62 (dd, 3J = 8.1 Hz, 4J = 2.1 Hz, 1H, C6H3), 4.24 (s, 2H, C5H4),
4.02 (s, 2H, C5H4), 3.97 (s, 2H, C5H4), 3.96 (s, 2H, C5H4), 2.75
(m, 2H, CH2), 2.34 (m, 2H, CH2) [minor isomer]
3b partially decomposed within 6 months. After purification by
chromatography, the ratio between the two isomers became 84 : 16.
7.79 (s, 2H, OH), 7.35–7.11 (m, 10H, 2C6H5), 6.74 (d, J = 8.0 Hz, 146.0, 145.8, 145.7, 144.9, 144.7, 142.9, 142.8, 141.3, 141.2,
1H, C6H3), 6.66 (d, J = 2.0 Hz, 1H, C6H3), 5.63 (dd, J = 8.0 and
2.0 Hz, 1H, C6H3), 4.15 (s, 5H, C5H5), 4.13 (s, 2H, CH2), 3.99
(t, J = 2.0 Hz, 2H, C5H4), 3.82 (t, J = 2.0 Hz, 2H, C5H4)
[major isomer]
3ca
137.4, 137.2, 133.7, 133.4 (Cquat), 130.7, 130.2, 129.2, 129.0,
128.7, 128.6, 127.2, 127.1, 126.4, 122.3, 122.0, 117.8, 117.5,
116.1, 116.0 (C6H5, C6H3), 87.8 (Fcipso), 70.7, 70.6 (C5H4),
70.0 (C5H5), 68.6 (C5H4), 41.3, 41.1 (CH2)
7.79 (s, 2H, OH), 7.35–7.11 (m, 10H, 2C6H5), 6.75 (d, J = 8.5 Hz,
1H, C6H3), 6.74 (d, J = 2.0 Hz, 1H, C6H3), 6.65 (dd, J = 8.5 and
2.0 Hz, 1H, C6H3), 4.16 (s, 5H, C5H5), 4.02 (s, 2H, CH2), 3.94
(t, J = 2.0 Hz, 2H, C5H4), 3.69 (t, J = 2.0 Hz, 2H, C5H4)
[minor isomer]
3c partially decomposed within 6 months. After purification by
chromatography, the ratio between the two isomers became 60 : 40.
7.43–6.43 (m, 13H, 2C6H5 + C6H3), 4.15 (s, 5H, C5H5), 4.06
(t, J = 1.9 Hz, 2H, C5H4), 3.54 (t, J = 1.9 Hz, 2H, C5H4)
[major isomer]
3da
145.0, 144.9, 142.4, 138.7, 135.1, 134.7, 132.9 (Cquat), 131.2,
130.1, 129.8, 128.8, 127.6, 127.4 (C6H5, C6H3), 87.7(Fcipso),
70.0 (C5H4), 69.5 (C5H5), 68.6 (C5H4)
7.43–6.43 (m, 13H, 2C6H5 + C6H3), 4.12 (s, 5H, C5H5), 4.01
(t, J = 1.9 Hz, 2H, C5H4), 3.39 (t, J = 1.9 Hz, 2H, C5H4)
[major isomer]
3d partially decomposed within 6 months. After purification by
chromatography, the ratio between the two isomers became 55 : 45
7.28–6.80 (m, 8H, C6H5 + C6H3), 4.06 and 4.04 (s, 5H, C5H5),
4.02 and 4.01 (t, J = 1.9 Hz, 2H, C5H4), 3.90 and 3.81 (t, J = 1.9
Hz, 2H, C5H4), 2.51 and 2.45 (q, J = 7.3 Hz, 2H, CH2), 2.16 (s,
3H, COCH3), 2.13 (s, 3H, COCH3), 0.98 and 0.92 (t, J = 7.3 Hz,
3H, CH3) [isomers present in 1 : 1 ratio]
4
168.2, 168.1 (COCH3), 143.9, 143.7, 143.1, 142.9, 141.7, 140.4,
138.9, 138.7, 136.1 (Cquat), 130.0, 129.5, 128.4, 128.3, 128.2,
127.5, 126.5, 124.8, 124.3, 123.1, 122.9 (C6H5, C6H3), 86.4
(Fcipso), 69.6, 69.5 (C5H4), 69.4, 69.3 (C5H5), 68.7, 68.4 (C5H4),
28.2, 27.9 (CH2), 20.7, 20.6 (COCH3), 15.5, 15.4 (CH3)
7546 | Dalton Trans., 2012, 41, 7537–7549
This journal is © The Royal Society of Chemistry 2012

View Article Online
Table 5 (Contd.)
Compound 1H NMR, δ (ppm)
13C NMR, δ/ppm
5/6
7.25–6.48 (m, 16H, C6H5 + C6H3), 5.19 and 5.14 (s, 1H, OH),
171.0 (COCH3), 148.4, 147.3, 147.2, 145.9, 145.7, 145.4, 145.3,
5.03 and 4.93 (s, 1H, OH), 4.05 (s, 5H, C5H5), 4.03 (s, 5H, C5H5), 145.0, 143.8, 140.2, 140.0, 139.5, 139.1, 138.6, 138.3 (Cquat),
4.02 and 3.99 (t, J = 1.9 Hz, 4H, C5H4), 3.88 and 3.79 (t, J = 1.9
131.5, 131.3, 130.9, 129.9, 129.7, 129.5, 128.5, 127.8, 127.6,
Hz, 4H, C5H4), 2.49 (m, 4H, CH2), 2.29 and 2.28 (s, 3H, COCH3), 125.5, 124.9, 124.0, 123.8, 123.7, 123.3, 120.5, 119.2, 119.0
2.26 and 2.22 (s, 3H, COCH3), 0.96 (m, 6H, CH3)
(C6H5, C6H3), 88.2, 88.1, 87.8, 87.7 (Fcipso), 70.9, 70.7 (C5H5),
69.9, 69.8, 69.7, 69.6 (C5H4), 29.5, 29.4 (CH2), 22.5 (COCH3),
17.2, 17.0 (CH3)
7a
7.67–7.41 (m, 10H, C6H5), 7.25–7.17 (m, 3H, CH, quinone ring),
187.2 (CvO), 157.3 (Cquat), 139.0 (C6H4, C6H5), 138.7 (Cquat),
6.36 (m, 1H, quinone ring), 6.24 (m, 1H, CH, vinyl), 4.44 (m, 1H, 137.8 (C6H4, C6H5), 137.4, 135.9 (Cquat), 131.3, 130.2 (C6H4,
C5H4), 4.37 (m, 1H, C5H4), 4.28 (m, 1H, C5H4), 4.23 (m, 1H,
C5H4), 4.00 (s, 5H, C5H5)
C6H5), 130.1 (Cquat), 129.4, 128.9, 128.8, 128.6, 128.5, 128.3,
127.5 (C6H4, C6H5), 87.8 (Fcipso), 69.7 (C5H5), 69.1, 69.0, 68.0,
66.6 (C5H4)
8aa
7.49–7.33 (m, 5H, C6H5), 6.64 (dd, J = 10.2 and 2.2 Hz, 1H,
182.0, 179.8 (CvO), 156.5, 150.5 (Cquat), 144.9 (C6H5, C6H3),
C6H3), 6.11 (dd, J = 10.2 and 0.6 Hz, 1H, C6H3), 5.83 (dd, J = 2.2 142.8, 136.3 (Cquat), 130.9, 129.8, 129.7, 128.3, 128.0 (C6H5,
and 0.6 Hz, 1H, C6H3), 4.58 (t, J = 1.8 Hz, 2H, C5H4), 4.53 (t, J = C6H3), 89.0 (Fcipso), 71.6, 71.3 (C5H4), 70.5 (C5H5), 44.7 (CH2),
1.8 Hz, 2H, C5H4), 4.28 (s, 5H, C5H5), 2.74 (q, J = 7.5 Hz, 2H,
CH2), 0.95 (t, J = 7.5 Hz, 3H, CH3) [major isomer]
15.7 (CH3)
7.49–7.33 (m, 5H, C6H5), 6.95 (dd, J = 10.1 and 2.2 Hz, 1H,
C6H3), 6.39 (dd, J = 2.2 and 0.6 Hz, 1H, C6H3), 6.32 (dd, J = 10.1
and 0.6 Hz, 1H, C6H3), 4.18 (t, J = 1.9 Hz, 2H, C5H4), 4.17 (s,
5H, C5H5), 3.91 (t, J = 1.9 Hz, 2H, C5H4), 2.92 (q, J = 7.5 Hz,
2H, CH2), 1.24 (t, J = 7.5 Hz, 3H, CH3) [minor isomer]
[isomers present in 81 : 19 ratio]
8ba
7.49–7.37 (m, 5H, C6H5), 6.95 (dd, J = 10.2 and 2.2 Hz, 1H,
181.6, 180.2 (CvO), 154.5, 145.5, 143.4, 142.8, 141.3 (Cquat),
C6H3), 6.03 (dd, J = 10.2 and 0.7 Hz, 1H, C6H3), 5.88 (dd, J = 2.2 131.5, 130.2, 129.6, 128.7, 127.7 (C6H5, C6H3), 88.2, 83.3
and 0.7 Hz, 1H, C6H3), 4.49 (t, J = 1.8 Hz, 2H, C5H4), 4.46 (t, J = (Fcipso), 71.7, 71.5, 70.1, 69.2 (C5H4), 41.8, 29.7 (CH2)
1.8 Hz, 2H, C5H4), 4.25 (t, J = 1.8 Hz, 2H, C5H4), 3.98 (t, J = 1.8
Hz, 2H, C5H4), 2.74 (m, 2H, CH2), 2.45 (m, 2H, CH2)
[major isomer]
7.49–7.37 (m, 5H, C6H5), 6.72 (dd, J = 10.1 and 2.2 Hz, 1H,
C6H3), 6.52 (dd, J = 2.2 and 0.8 Hz, 1H, C6H3), 6.28 (dd, J = 10.1
and 0.8 Hz, 1H, C6H3), 4.37 (t, J = 1.8 Hz, 2H, C5H4), 4.07 (t, J =
1.8 Hz, 2H, C5H4), 4.04 (t, J = 1.8 Hz, 2H, C5H4), 3.94 (t, J = 1.8
Hz, 2H, C5H4), 3.05 (m, 2H, CH2), 2.51 (m, 2H, CH2)
[minor isomer] [isomers present in 84 : 16 ratio]
8ca
7.38–7.11 (m, 10H, 2C6H5), 6.72 (dd, J = 10.2 and 2.2 Hz, 1H,
The instability of this compound precluded characterization by
C6H3), 6.14 (dd, J = 10.2 and 0.7 Hz, 1H, C6H3), 5.92 (dd, J = 2.2 13C NMR.
and 0.7 Hz, 1H, C6H3), 4.45 (broad singlet, 4H, tentatively
attributed to C5H4), 4.35 (s, 5H, C5H5), 4.13 (s, 2H, CH2)
[major isomer]
7.38–7.11 (m, 10H, 2C6H5), 6.64 (dd, J = 10.1 and 2.1 Hz, 1H,
C6H3), 6.45 (dd, J = 2.1 and 0.6 Hz, 1H, C6H3), 6.06 (dd, J = 10.1
and 0.6 Hz, 1H, C6H3), 4.24 (s, 5H, C5H5), 4.15 (t, J = 1.9 Hz,
2H, C5H4), 3.88 (t, J = 1.9 Hz, 2H, C5H4), the signal for CH2Ph
could not be clearly identified [minor isomer] [isomers present in
65 : 35 ratio]
7.69–7.10 (m, 10H, 2C6H5), 6.83 (dd, J = 10.2 and 2.2 Hz, 1H,
C6H3), 5.95 (dd, J = 2.2 and 0.6 Hz, 1H, C6H3), 5.88 (dd, J = 2.2
and 0.6 Hz, 1H, C6H3), 4.25 (s, 5H, C5H5), 4.24 (t, J = 2.0 Hz,
2H, C5H4), 3.51 (t, J = 2.0 Hz, 2H, C5H4) [major isomer]
7.69–7.10 (m, 10H, 2C6H5), 6.82 (dd, J = 10.1 and 2.2 Hz, 1H,
C6H3), 6.29 (dd, J = 10.1 and 0.7 Hz, 1H, C6H3), 6.23 (dd, J = 2.2
and 0.7 Hz, 1H, C6H3), 4.48 (t, J = 1.9 Hz, 2H, C5H4), 4.30 (t, J =
1.9 Hz, 2H, C5H4), 4.28 (s, 5H, C5H5) [minor isomer]
[isomers present in 73 : 27 ratio]
8da
181.4, 180.0 (CvO), 155.3, 147.2, 142.9, 142.0, 136.4 (Cquat),
131.7, 131.6, 131.1, 130.0, 128.9, 128.7, 127.7 (C6H5, C6H3),
84.7 (Fcipso), 70.6 (C5H5), 71.6, 70.7 (C5H4)
a In CD3COCD3.
with 5 mL of acetonitrile containing Bu4NBF4 (0.2 mol L−1).
Compounds 3a or 3c were introduced in the anodic compartment
and 9-fluorenone (0.25 mmol) was introduced in the cathodic
compartment in order to provide a reproducible and easy auxili-
ary cathodic reaction. The electrolyses were stopped until a
charge of 2 Faraday per mole was passed. The electrolyses were
monitored by cyclic voltammetry (CV) at a 0.5 mm Pt electrode.
Electrolyses and CVs were recorded on a PAR 2273 potentiostat.
3.6 Biochemistry
The human cell MDA-MB-231 (breast adenocarcinoma) pur-
chased from ATCC was grown in RPMI medium supplemented
with 10% fetal calf serum, in the presence of penicillin, strepto-
mycin and fungizone in 75 cm3 flasks under 5% CO2. For cyto-
toxicity determinations, cells were plated in 96-well tissue
culture microplates in 200 μL complete medium and treated 24 h
This journal is © The Royal Society of Chemistry 2012
Dalton Trans., 2012, 41, 7537–7549 | 7547

View Article Online
Table 6 MS and analytical data for compounds 2c–e and 3–7
MS, m/z
HRMS, m/z
Compound
Found (calculated for M+)
Found (calculated for M+)
Elemental analysis: found (calculated) %
2c
2d
2e
3a
3b
3c
3d
4
470 (470)
456 (456)
487 [MH+] (486)
424 (424)
422 (422)
486 (486)
472 (472)
508 (508)
466 (466)
468 (468)
C31H26FeO: 470.1327 (470.1333)
C30H24FeO: 456.1171 (456.1177)
C31H26FeO2: 486.1277 (486.1282)
C26H24FeO2: 424.1120 (424.1126)
C26H22FeO2: 422.0958 (422.0969)
C31H26FeO2: 486.1282 (486.1282)
C30H24FeO2: 472.1125 (472.1126)
C30H28FeO4: 508.1332 (508.1337)
C28H26FeO3: 466.1221 (466.1231)
C31H24FeO: 468.1171 (468.1177)
C31H26FeO·0.25H2O: C, 78.57 (78.40); H, 5.68 (5.62)a
C30H24FeO·0.5H2O: C, 77.39 (77.43); H, 5.43 (5.41)a
C31H26FeO2·H2O: C, 74.09(73.82); H, 5.89 (5.60)a
C26H24FeO2: C, 73.19 (73.59); H, 5.79 (5.70)
C26H22FeO2·0.25H2O: C, 72.87 (73.12); H, 5.35 (5.31)a
C31H26FeO2·0.5H2O: C, 75.18 (75.16); H, 5.62 (5.49)a
C30H24FeO2·0.25H2O: C, 75.45 (75.56); H, 5.52 (5.18)a
C30H28FeO4·0.5H2O: C, 69.61 (69.94); H, 5.61 (5.65)a
C28H26FeO3·0.5C6H14: C, 72.94 (73.09); H, 6.59 (6.53)b
C31H24FeO·2H2O: C, 74.11 (73.82); H, 5.91 (5.60)a
5/6
7
1
1
a Presence of water in the analytical sample was verified by H NMR spectroscopy. b Presence of hexane in the analytical sample was verified by H
NMR spectroscopy.
Table 7 Crystal data for E-2c, Z-2d, Z-2e and Z-4
Compound
E-2c
Z-2d
Z-2e
Z-4
Formula
Fw
C31H26FeO
470.37
C30H24FeO
456.34
C31H26FeO2
486.37
C30H28FeO4
508.37
Crystal system
Space group
Monoclinic
P21/c
Monoclinic
P21/n
Triclinic
P1
Monoclinic
P21/c
ˉ
Unit cell dimensions
a (Å)
b (Å)
c (Å)
α (°)
14.950(2)
9.9214(13)
15.579(2)
90
98.413(7)
90
2285.8(5)
4
1.367
9.7144(4)
12.9617(5)
17.6055(7)
90
95.1240(10)
90
2207.95(15)
4
1.373
9.4498(2)
12.1301(4)
17.0739(5)
12.8231(3)
90
112.2770(10)
90
2457.55(12)
4
1.374
11.5084(2)
12.0193(2)
71.0240(10)
81.4350(10)
66.1580(10)
1130.37(4)
2
1.429
0.695
508
0.40 × 0.40 × 0.06
1.79 to 31.07
19 666
4609 (0.0309)
99.8 (26.37)
0.9595–0.7685
4609/1/321
1.195
R1 = 0.0322,
wR2 = 0.0924
R1 = 0.0410,
wR2 = 0.1180
0.426 and −0.482
β (°)
γ (°)
Volume (Å3)
Z
ρc (mg m−3
)
μ (Mo Kα) (mm−1
F(000)
)
0.682
984
0.703
952
0.648
1064
Crystal size (mm3)
0.40 × 0.20 × 0.10
2.44 to 30.51
49 324
4676 (0.0650)
99.9 (26.37)
0.9350–0.7722
4676/6/308
1.220
R1 = 0.0441,
wR2 = 0.1159
R1 = 0.0773,
wR2 = 0.1543
0.564 and −0.617
0.33 × 0.11 × 0.11
2.31 to 26.36
26 025
4461 (0.0530)
99.0 (26.36)
0.9266–0.8011
4461/0/303
1.166
R1 = 0.0464,
wR2 = 0.1516
R1 = 0.0573,
wR2 = 0.1676
0.484 and −0.553
0.40 × 0.31 × 0.22
2.09 to 26.37
56 571
5015 (0.0543)
99.8 (26.37)
0.8705–0.7815
5015/0/428
1.168
R1 = 0.0331,
wR2 = 0.1036
R1 = 0.0422,
wR2 = 0.1252
0.611 and −0.661
θ range (°)
Reflections collected
Independent reflections (Rint
Completeness % (to θ, deg)
Transmission range
Data/restraints/parameters
Goodness-of-fit on F2
Final R indices [I > 2σ(I)]
)
R indices (all data)
Largest diff. peak and hole (e Å−3
)
later with compounds dissolved in DMSO using a Biomek 3000
(Beckman-Coulter). Controls received the same volume of
DMSO (1% final volume). After 72 h exposure, MTS reagent
(Promega) was added and incubated for 3 h at 37 °C: the absor-
bance was monitored at 490 nm and results expressed as the inhi-
bition of cell proliferation calculated as the ratio {[1 − (OD490
treated/OD490 control)] × 100} in triplicate experiments. For
IC50 experiments performed in duplicate, compounds were
added in the range 0.5 nM–10 μM in a fixed volume of DMSO.
radiation, with the SMART suite of programs.35 Data were pro-
cessed and corrected for Lorentz and polarization effects with
SAINT,36 and for absorption effects with SADABS.37 Structural
solution and refinement were carried out with the SHELXTL
suite of programs.38 Crystal and refinement data are summarized
in Table 7.
The structures were solved by direct methods to locate the
heavy atoms, followed by difference maps to complete the
structure. Organic hydrogen atoms were placed in calculated pos-
itions and refined with a riding model; hydroxyl hydrogen atoms
were either located or placed in calculated positions and refined
with appropriate distance restraints. All non-hydrogen atoms
were given anisotropic displacement parameters in the final
model.
3.7 X-ray crystal structure determinations
Crystals were mounted on quartz fibers. X-ray data were col-
lected using a Bruker AXS APEX system, using Mo Kα
7548 | Dalton Trans., 2012, 41, 7537–7549
This journal is © The Royal Society of Chemistry 2012

View Article Online
9 G. K. Poon, B. Walter, P. E. Lonning, M. N. Holton and R. McCague,
Drug Metab. Dispos., 1995, 23, 377–382.
10 X. Liu, E. Pisha, D. A. Tonetti, D. Yao, Y. Li, J. Yao, J. E. Burdette and
J. L. Bolton, Chem. Res. Toxicol., 2003, 16, 832–837.
11 S. S. Dehal and D. Kupfer, Cancer Res., 1995, 56, 1283–1290.
12 S. Top, J. Tang, A. Vessières, D. Carrez, C. Provot and G. Jaouen, Chem.
Commun., 1996, 955–956.
There was disorder of the hydroxyl group in E-2c and Z-2e
over two aromatic rings in each case. These were refined with
appropriate restraints and on their occupancies which were con-
strained to sum to unity.
13 S. Top, A. Vessières, C. Cabestaing, I. Laios, G. Leclercq, C. Provot and
G. Jaouen, J. Organomet. Chem., 2001, 637–639, 500–506.
14 S. Top, A. Vessières, G. Leclercq, J. Quivy, J. Tang, J. Vaissermann,
M. Huché and G. Jaouen, Chem.–Eur. J., 2003, 9, 5223–5236.
15 G. Jaouen, S. Top, A. Vessières, G. Leclercq and M. J. McGlinchey,
Curr. Med. Chem., 2004, 11, 2505–2517.
16 A. Vessières, S. Top, W. Beck, E. A. Hillard and G. Jaouen, Dalton
Trans., 2006, 529–541.
17 E. A. Hillard, P. Pigeon, A. Vessières, C. Amatore and G. Jaouen, Dalton
Trans., 2007, 5073–5081.
Conclusion
In our ongoing investigations on the anti-proliferative activity of
conjugated phenylferrocene systems, we have synthesized a
series of ferrocenyl catechols via the McMurry coupling reaction
and tested their anti-proliferative effects on hormone-indepen-
dent breast cancer cell lines. These ferrocenyl catechols 3 were
found to have equivalent or higher efficacy than their phenolic
analogues 2, in contrast to previous reports on the phenol and
catechol metabolites of tamoxifen and toremifen. The ability to
form oxidation products, either QMs or OQs is closely related to
their activity in the in vitro assays. Consistent with our earlier
investigations, the catechol bearing the [3]ferrocenophane motif,
viz. 3b, displays greater anti-proliferative potency than its ferro-
cenyl analogue 2b.
Interestingly, molecules that could theoretically form QM or
OQs (3a and 3c) seem to exclusively favor the OQ structure
under chemical oxidation. We were surprised that, whatever the
metabolites favored in the biological milieu, the anti-proliferative
effects are quite similar. Further work will include metabolism
studies using biologically relevant oxidants. The reaction of QM
and OQ forms of these ferrocenyl phenol and catechol com-
pounds with biologically relevant nucleophiles such as gluta-
thione and deoxynucleosides are the subject of current
investigations.
18 E. A. Hillard, A. Vessières, F. Le Bideau, D. Plażuk, D. Spera, M. Huché
and G. Jaouen, ChemMedChem, 2006, 1, 551–559.
19 A. Valentini, F. Conforti, A. Crispini, A. D. Martino, R. Condello,
C. Stellitano, G. Rotilio, M. Ghedini, G. Federici, S. Bernardini and
D. Pucci, J. Med. Chem., 2009, 52, 484–491.
20 E. A. Hillard, A. Vessières, L. Thouin, G. Jaouen and C. Amatore,
Angew. Chem., Int. Ed., 2006, 45, 285–290.
21 D. Hamels, P. M. Dansette, E. A. Hillard, S. Top, A. Vessières, G. Jaouen
and D. Mansuy, Angew. Chem., Int. Ed., 2009, 48, 9124–9126.
22 D. Plazuk, A. Vessières, E. A. Hillard, O. Buriez, E. Labbé, P. Pigeon,
M.-A. Plamont, C. Amatore, J. Zakrzewski and G. Jaouen, J. Med.
Chem., 2009, 52, 4964–4967.
23 M. Gormen, D. Plazuk, P. Pigeon, E. A. Hillard, M.-A. Plamont, S. Top,
A. Vessières and G. Jaouen, Tetrahedron Lett., 2010, 51, 118–120.
24 M. Gormen, P. Pigeon, S. Top, A. Vessières, M.-A. Plamont, E.
A. Hillard and G. Jaouen, Med. Chem. Commun., 2010, 1, 149–151.
25 During the preparation of this manuscript, we discovered the article
F. Zhao, C. Zhao and Z.-Q. Liu, J. Biol. Inorg. Chem, 2011, 16, 1169–
1176. This article reports the synthesis of 2d, 3a and 3d by similar
methods as reported here.
26 Y. L. K. Tan, P. Pigeon, E. A. Hillard, S. Top, M.-A. Plamont,
A. Vessières, M. J. McGlinchey, H. Müller-Bunz and G. Jaouen, Dalton
Trans., 2009, 10871–10881.
27 P. Messina, E. Labbé, O. Buriez, E. A. Hillard, A. Vessières, D. Hamels,
S. Top, G. Jaouen, Y. M. Frapart, D. Mansuy and C. Amatore, Chem.–
Eur. J., 2012, DOI: 10.1002/chem.201103378.
Acknowledgements
28 A. Vessières, S. Top, P. Pigeon, L. Boubeker, D. Spera and G. Jaouen,
J. Med. Chem., 2005, 48, 3937–3940.
29 J. B. Heilmann, E. A. Hillard, M. A. Plamont, P. Pigeon, M. Bolte,
G. Jaouen and A. Vessières, J. Organomet. Chem., 2008, 693, 1716–
1722.
We thank Dr Li Yongxin for crystallographic assistance and
Dr Thierry Cresteil for biochemical tests. Financial support
was provided by the Agence Nationale de la Recherche
(No. ANR-06-BLAN-0384-01, “FerVect”, postdoctoral fellow-
ship for Y. L. K. T.) and COST (D39 WG001).
30 D. Yao, F. Zhang, L. Yu, Y. Yang, R. B. van Breemen and J. L. Bolton,
Chem. Res. Toxicol., 2001, 14, 1643–1653.
31 M. Gormen, P. Pigeon, S. Top, E. A. Hillard, M. Huché, C. G. Hartinger,
F. de Montigny, M.-A. Plamont, A. Vessières and G. Jaouen, ChemMed-
Chem, 2010, 5, 2039–2050.
32 D. Enders, R. Peters, G. Raabe, J. Runsick and J. W. Bats, Eur. J. Org.
Chem., 2000, 3399–3426.
33 A. C. Benyei, C. Glidewell, P. Lightfoot, B. J. L. Royles and D.
M. Smith, J. Organomet. Chem., 1997, 539, 177–186.
34 T. D. Turbitt and W. E. Watts, J. Organomet. Chem., 1972, 46, 109–117.
35 SMART, version 5.628, Bruker Advanced Analytical X-ray Systems, Inc.,
Madison, WI, USA, 2001.
References
1 B. Fisher, J. Costantino, C. Redmond, R. Poisson, D. Bowman,
J. Couture, N. V. Dimitrov, N. Wolmark, D. L. Wickerham, E. R. Fisher,
R. Margolese, A. Robidocex, H. Shibotu, J. Terz, A. H. G. Paterson, H.
L. Lickley and M. Ketner, N. Engl. J. Med., 1989, 320, 479–484.
2 C. K. Osbourne, N. Engl. J. Med., 1998, 339, 1609–1618.
3 G. N. Hortobagyi, N. Engl. J. Med., 1998, 339, 974–984.
4 S. G. Nayfield, J. E. Karp, L. G. Ford, F. A. Dorr and B. S. Kramer,
J. Nat. Cancer Inst., 1991, 23, 1450–1459.
36 SAINT+, version 6.22a, Bruker Advanced Analytical X-ray Systems,
Inc., Madison, WI, USA, 2001.
37 G. M. Sheldrick, SADABS, Program for area detector adsorption correc-
tion, Institute for Inorganic Chemistry, University of Göttingen, Germany,
1996.
38 SHELXTL, version 5.1, Bruker Advanced Analytical X-ray Systems, Inc.,
Madison, WI, USA, 1997.
5 V. C. Jordan, J. Med. Chem., 2003, 46, 883–908.
6 V. C. Jordan, J. Med. Chem., 2003, 46, 1081–1108.
7 L. L. Smith and I. N. White, Oncology, 1998, 12, 14–22.
8 M. A. Killackey, T. B. Hakes and V. K. Price, Cancer Treat. Rep., 1985,
69, 237–238.
This journal is © The Royal Society of Chemistry 2012
Dalton Trans., 2012, 41, 7537–7549 | 7549