Organometallic cluster analogues of tamoxifen: Synthesis and biochemical assay

Article abstract of DOI:10.1016/j.jorganchem.2005.08.041

Simple organometallic cluster analogues of tamoxifen containing triosmium or dicobalt carbonyl fragments have been prepared. Attempts at elaboration of these towards the tamoxifen skeleton were hampered by sensitivity of the cluster-ligand linkage towards

Full text of DOI:10.1016/j.jorganchem.2005.08.041

Journal of Organometallic Chemistry 691 (2006) 9–19
www.elsevier.com/locate/jorganchem
Organometallic cluster analogues of tamoxifen: Synthesis
and biochemical assay
a
a,
b
b
Kiat Hwa Chan , Weng Kee Leong *, Gerard Jaouen , Laurence Leclerq ,
b
b
`
Siden Top , Anne Vessieres
a
Department of Chemistry, National University of Singapore, 3 Science Drive3, Singapore 117543, Singapore
Laboratoire de Chimie et Biochimie des Complexes Mole´culaires, Ecole Nationale Supe´rieure de Chimie de Paris, UMR C.N.R.S. 7576,
11, rue Pierre et Marie Curie, 75231 Paris Cedex 05, France
b
Received 12 August 2005; accepted 26 August 2005
Available online 4 October 2005
Abstract
Simple organometallic cluster analogues of tamoxifen containing triosmium or dicobalt carbonyl fragments have been prepared.
Attempts at elaboration of these towards the tamoxifen skeleton were hampered by sensitivity of the cluster–ligand linkage towards
the McMurry coupling conditions. Preliminary biological tests on various substrates indicate that the transition metal carbonyl fragment
increases lipophilicity dramatically and reduces affinity for the estrogen receptor.
ꢀ 2005 Elsevier B.V. All rights reserved.
Keywords: Osmium; Cobalt; Cluster; Tamoxifen
1. Introduction
attributed to the cytotoxic activity of the ferrocenium
ion, which results from the oxidation of the ferrocene moi-
ety in vivo [6e].
Tamoxifen is widely used for the treatment of hormone-
dependent breast cancer [1]. However, it is known to be
completely effective for a long period only in about one
in three such cases. There is thus still considerable effort
put into the development of modified tamoxifen, either to
increase its efficacy directly, or to use it as a vehicle for
introducing other cytotoxic agents. In addition to organic
modifications, inorganic and organometallic modifications
have been explored as well. These include platinum com-
plexes [2], carborane complexes [3], organorhenium [4],
titanocene [5], ferrocene [6], and ruthenocene [7]. Among
these, the most well-studied and potentially most successful
metal-containing analogues of tamoxifen are those con-
taining the ferrocene moiety, or ferrocifens. For example,
hydroxyferrocifen was found to have a stronger anti-prolif-
erative effect than hydroxytamoxifen, and this has been
Hydroxyferrocifen was originally synthesized via a series
of Grignard and dehydration reactions [6a], but since then
this has been superseded by a more efficient McMurry cou-
pling route (Scheme 1) [6e]. The same McMurry coupling
route has been adopted for the incorporation of CpRe(CO)3
into the tamoxifen skeleton, illustrating the usefulness of
the McMurry protocol [4a]. Since tamoxifen is capable of
tolerating a considerable degree of structural variation
while still retaining its affinity for the estrogen receptor,
we were interested in investigating the effect of incorporat-
ing a metal carbonyl cluster. It was hoped that the multiple
low-valent metal centres in such derivatives may be oxidized
to yield a multiply-charged cluster which would represent a
potent cytotoxic species contained within a small volume,
enabling tumour suppression to be more effective.
No organometallic cluster-containing derivative of
tamoxifen has ever been reported. We were thus interested
in exploring the possibilities for synthesizing such a class of
*
Corresponding author.
E-mail address: chmlwk@nus.edu.sg (W.K. Leong).
0022-328X/$ - see front matter ꢀ 2005 Elsevier B.V. All rights reserved.
doi:10.1016/j.jorganchem.2005.08.041

10
K.H. Chan et al. / Journal of Organometallic Chemistry 691 (2006) 9–19
OH
OH
Zn, TiCl /THF
4
O
O
reflux
(McMurry
reagent)
Fe
Fe
OH
OH
NaH, acetone
ClCH CH NMe ,HCl
2
2
2
OH
Fe
NMe
2
O
Scheme 1.
modified tamoxifens. Our plan was to explore the synthetic
routes towards a tamoxifen-like derivative containing a tri-
osmium carbonyl cluster, as these clusters are generally
quite stable. In this paper, we wish to report our attempts
at this goal.
OH
2. Results and discussion
O
2.1. Studies on the incorporation of clusters into tamoxifen
O
Os
Our initial explorations involved tamoxifen-like deriva-
tives containing a carboxylate-cluster linkage as these link-
ages are expected to be stable, and hence would also be
amenable to further transformations of the cluster; we were
thus interested in derivatives such as that given in Fig. 1.
For the synthesis of triosmium-carboxylate clusters, the
parent triosmium carbonyl cluster Os3(CO)12, 1a, is of lim-
ited utility as it is rather inert; often an activated derivative
such as Os3(CO)10(NCMe), 2a, Os3(CO)10(NCMe)2, 2b, or
Os3(CO)10(l-H)(l-OH), 3, is employed instead. Trios-
mium-carboxylate clusters can be conveniently prepared
by the reaction of 3 with the appropriate carboxylic acid
in the presence of HBF4 Æ OEt2 as an acid catalyst [8]. We
have thus prepared a series of such triosmium-carboxylate
clusters, 5, starting from the appropriate carboxylic acids,
4 (Scheme 2 and Table 1). Clusters 5a–f were simple but
close analogues to the target; the alkenic functionality in
Os
Os
H
OH
Fig. 1. Targetted triosmium cluster derivative of tamoxifen.
tamoxifen was retained, and they were synthesized for tests
on their potential biochemical activity.
These reactions leading to the carboxylate clusters can
be quite fast when acid is present; thus, for example, H
NMR spectroscopic monitoring of the reaction of 3 and
crotonic acid, 4c, showed the instantaneous formation of
E-Os3(CO)10(l-H){l-O2C@C(H)C(H)CH3}, 5e. In the ab-
sence of HBF4, no reaction occurred even after 6 h. This
is in contrast to 2b, which reacted gradually with carboxylic
acids [9]. However, the ease of reaction was found to
depend on the size and functional groups present on the
carboxylic acids. While the reactions of smaller carboxylic
1

K.H. Chan et al. / Journal of Organometallic Chemistry 691 (2006) 9–19
11
HCl in PhCH3, CH2Cl2, THF, CH3CN, and DMSO, since
HCl was a likely side-product from hydrolysis of the
McMurry reagent; it remained stable in all the solvents
tested except CH3CN, in which it was converted to the
known Os3(CO)10(l-H)(l-Cl), 6. The formation of 6 in
CH3CN could be due to labilization of the carboxylate li-
gand by CH3CN itself; for example, nucleophilic substitu-
tion in clusters of the type Os3(CO)10(l-H)(l-OR) by EPh3
(E = P, As, Sb) occur at a bridged Os, suggesting that it is
the electrophilic site [11]. A similar process involving initial
attack by CH3CN could have occurred.
Unfortunately, the attempted McMurry coupling of 5g
and 5h with 4,40-dihydroxybenzophenone both yielded
cluster 6 instead of the McMurry coupled product. The
identity of 6 from this reaction has also been confirmed
by a single crystal X-ray crystallographic study (see Sup-
porting Information), which yielded a different polymorph
from that reported earlier [12]. Since we have shown that
the carboxylate linkage itself was stable towards the
McMurry reagent, the substitution of RCOOꢀ by Clꢀ
was presumably related to the presence of the a-keto
group, which could have labilized the carboxylate linkage
through chelation to the McMurray reagent (Fig. 3). At-
tempts to circumvent this by moving the keto group further
away and yet retaining the proximity of the cluster moiety
to the two phenolic rings (the sites of estrogen receptor rec-
ognition) [13], viz., 5i and 5j, were also unsuccessful, lead-
ing again to the formation of 6. The McMurry reagent has
been reported to be capable of forming alkenic macrocycles
with ring sizes from 4 to 23 [14], and this may be operative
here. Hopes that the inflexible phenyl separating the keto
and carboxylate groups in 5k may resolve the problem were
unfortunately pre-empted by its low yield.
R
O
O
Os
HBF4.OEt2
RT
3
+
RCOOH
Os
Os
4
H
5
Yield(%)
41
a
b
c
d
e
f
CH=CH2
CEt=CH2
86
E-CH=CHMe
E-CH=CHPh
82
77
E-CH=CH-p-C6H4OH
E-CMe=CHPh
COEt
0.4
80
g
h
i
66
COPh
72
CH2CH2COMe
E-CH=CHCOPh
p-C6H4COMe
97
j
89
k
1.2
Scheme 2.
acids (e.g., 4c and 4g) were completed in 10 min, larger
carboxylic acids (e.g., 4i and 4j) required more than 5 h
to reach completion. On the other hand, while 4d and 4i
gave good yields of 5d (77%) and 5i (97%), respectively,
the closely related acids 4e and 4k gave very poor yields
of 5e (0.4%) and 5k (1.2%), respectively. These point to
subtle electronic effects; slightly better yields of clusters 5e
and 5k were obtained when the corresponding carboxylic
acids were reacted with 2b (14% and 5.3%, respectively).
The structure of 5h has also been confirmed by a single
crystal X-ray crystallographic study. There are two crystal-
lographically independent molecules in the asymmetric
unit; the ORTEP diagram for one of the molecules, to-
gether with selected bond parameters, is shown in Fig. 2.
The molecular structure shows that the benzoyl formate li-
gand occupies two axial positions on adjacent osmium
As it appeared that the problem was with the oxophilic-
ity of titanium, we thought that perhaps the chelation effect
may be diminished if the carboxylate linkage was replaced
by a thiolate; the softer sulfur was also expected to bind
more strongly with the low-valent osmium. Two thiolate
clusters were prepared using the same procedure as for
the carboxylate clusters, viz., Os3(CO)10(l-H){l-
SC(@O)Ph}, 5l, and Os3(CO)10(l-H)(l-SCH2CH2OH),
5m, and tested in their reactions with the McMurry re-
agent. Cluster 5l has previously been synthesized from
the reaction of thiobenzoic acid with Os3(CO)10(MeCN)2
[15]. Although 5m was inert to the McMurry reagent, as
˚
˚
atoms. The Os(1A)–Os(2A) [2.9297(5) A] edge is longer
than the Os(1A)–Os(3A) [2.8562(4) A] and Os(2A)–
Os(3A) [2.8708(5) A] edges, which is a common feature of
˚
hydride–bridged Os–Os edges [10]. The Os(1A)–C(12A)
1
˚
˚
[1.867(6) A] and Os(2A)–C(22A) [1.864(6) A] bonds are
judged by H NMR monitoring, the reaction of 5l and
4,40-dihydroxybenzophenone with the McMurry reagent
led to the known cluster Os3(CO)9(l-H)2(l3-S), 7, instead.
Thus it appeared that there was nucleophilic attack by Clꢀ
on the carbonyl group of 5n leading to displacement of the
cluster moiety; this was supported by the observation that 7
was formed when 5l was treated with aq. HCl (Scheme 3).
An attempt to circumvent the nucleophilicity of the
chloride ion by adding AgBF4 to remove it from the
McMurry reagent prior to the addition of 5g and 4,40-
dihydroxybenzophenone led to an intractable brown
mixture.
˚
shorter than the Os(3A)–C(32A) [1.964(7) A] bond. This re-
flects the r-donating and weak p-accepting properties of the
carboxylate ligand, which together results in more electron
density being donated from Os(1A) and Os(2A) to the trans
carbonyls, CO(12A) and CO(22A), respectively.
Since the McMurry reaction involved Lewis acidic con-
ditions, cluster 5c was first tested under the normal
McMurry conditions (reflux with the McMurry reagent);
5c was detected as the only hydride-containing cluster,
indicating that the carboxylate linkage would be stable in
the McMurry reaction. Cluster 5h was also reacted with

12
K.H. Chan et al. / Journal of Organometallic Chemistry 691 (2006) 9–19
Table 1
Spectroscopic data for cluster compounds 5
Compound
m(CO)/cmꢀ1
1H NMR/d
Elemental analysis: Calc. (Found)/%
3
5a
2114w, 2075vs, 2064s,
2028vs, 2017vs, 2011w,
1988w
5.97 (dd, 1H, CH@CH2, Jtrans = 17.32 Hz,
2Jgem = 2.47 Hz), 5.87 (dd, 1Ha,
For C13H4O12Os3: C,16.92; H, 0.44.
Found: C, 17.06; H, 0.30
CH@CH2,3Jcis = 9.90 Hz), 5.15 (dd, 1H,
CH@CH2), ꢀ10.42 (s, 1H, OsHOs)
5b
5c
2113w, 2074vs, 2064s,
2028vs, 2017vs, 2012w,
1983w
2113w, 2075vs, 2064s,
2027vs, 2017vs, 2010w,
1984w
5.76 (s, 1H, C@CH2), 5.25 (s, 1H, C@CH2), 2.05
(q, 2Hb, CH2CH3, 3J = 7.43 Hz), 0.92 (t, 3H,
CH2CH3), ꢀ10.42 (s, 1H, OsHOs)
For C15H8O12Os3: C, 18.95; H, 0.85.
Found: C,18.90; H, 0.64
3
6.56 (dq, 1H, C@CHCH3, Jcis = 6.60 Hz,
For C14H6O12Os3: C, 17.95; H, 0.65.
Found: C, 18.31; H, 0.95
3Jtrans = 15.26 Hz), 5.62 (dq, 1H, CH@CHCH3,
4J = 1.65 Hz), 1.80 (dd, 3H, CH@CHCH3),
ꢀ10.43 (s, 1H, OsHOs)
5d
5e
2113w, 2075vs, 2063s,
2027vs, 2017vs, 2010w,
1985w
2113w, 2075vs, 2063s,
2027vs, 2017vs, 2010w,
1985w
7.44–7.40 (m, 2H, o-C6H5), 7.35–7.33 (m, 3H, m,p-
C6H5), 7.24 (d, 1Hb, CH@CH,3Jtrans = 16.50 Hz),
6.23 (d, 1H, CH@CH), ꢀ10.35 (s, 1H, OsHOs)
For C19H8O12Os3 Æ 1/4C6H14: C, 24.13;
H, 1.14. Found: C, 24.35; H, 1.06
3
7.32 (d, 2H, C6H4, Jcd = 9.08 Hz), 7.18 (d, 1H,
For C19H8O13Os3 Æ 1/4C6H14: C, 23.75;
H, 1.12. Found: C, 23.46; H, 0.85
CH@CH,3Jtrans = 15.67 Hz), 6.79 (d, 2H, C6H4),
6.09 (d, 1H, CH@CH), 4.96 (s, 1H, OH), ꢀ10.35
(s, 1H, OsHOs)
5f
5g
5h
5i
2113w, 2075vs, 2064s,
2027vs, 2017vs, 2013w,
1982w
2113w, 2074vs, 2063s,
2025vs, 2017vs, 2010w,
1986w
2114w, 2075vs, 2063s,
2027vs, 2017vs, 2010w,
1984w
2113w, 2074vs, 2064s,
2027vs, 2017vs, 2012w,
1986w
7.37–7.24 (m, 6H, C6H5 and C@CH), 1.85 (d, 3H,
For C20H10O12Os3: C, 23.71; H, 1.00.
Found: C, 23.65; H, 0.83
4
CH3C@C, Jtrans = 1.85 Hz), ꢀ10.27 (s, 1H,
OsHOs)a
2.59 (q, 2H, CH2CH3, 3J = 7.43 Hz), 0.99 (t, 3H,
CH2CH3), ꢀ10.45 (s, 1H, OsHOs)
For C14H6O13Os3: C, 17.65; H, 0.64.
Found: C, 17.93; H, 0.83
7.69–7.61 (m, 3H, C6H5), 7.52–7.46 (m, 2H,
C6H5), ꢀ10.25 (s, 1H, OsHOs)
For C18H6O13Os3: C, 21.60; H, 0.60.
Found: C, 21.62; H, 0.55
2.52 (t, 2H, CH2CH2, 3J = 5.78 Hz), 2.45 (t, 2H,
CH2CH2), 2.13 (s, 3H, COCH3), ꢀ10.45 (s, 1H,
OsHOs)
For C15H8O13Os3: C, 18.64; H, 0.84.
Found: C, 18.86; H, 1.05
3
5j
2113w, 2075vs, 2062s,
2028vs, 2017vs, 2010w,
1987w
7.90 (d, 2H, o-C6H5, Jcd = 8.25 Hz), 7.61 (t, 1H,
For C20H8O13Os3: C, 23.39; H, 0.79.
Found: C, 23.49; H, 0.80
3
p-C6H5, Jde = 7.42 Hz), 7.50 (td, 2H, m-C6H5),
7.41 (d, 1H, CH@CH,3Jtrans = 15.67 Hz), 6.60 (d,
1H, CH@CH), ꢀ10.38 (s, 1H, OsHOs)
7.88 (d, 2H, C6H4, 3J = 8.25 Hz), 7.78 (d, 2H,
C6H4), 2.59 (s, 3H, COCH3), ꢀ10.25 (s, 1H,
OsHOs)
5k
5m
a
2114w, 2075vs, 2064s,
2027vs, 2017vs, 2011w,
1986w
2110w, 2069vs, 2059s,
2025vs, 2021sh, 2001w,
1990w, 1984w
For C19H8O13Os3: C, 22.48; H, 0.79.
Found: C, 22.80; H, 0.52
3.84 (td, 2H, CH2OH), 2.61 (t, 2H, SCH2,
3J = 5.79 Hz), 1.74 (t, 1H, OH, 3J = 4.95 Hz),
ꢀ17.36 (s, 1H)
For C12H6O11Os3S: C, 15.52; H, 0.65; S,
3.45. Found: C, 15.51; H, 0.79; S, 3.01
In CD2Cl2.
Several unsuccessful attempts at incorporating a func-
tionality such as a carboxylic acid, phenol, trichloromethyl,
nitrile or phosphine, which can be subsequently functional-
ized with an organometallic cluster, onto the tamoxifen
skeleton were also made. A partially successful strategy
was found with the incorporation of an alkyne functional-
ity following a similar procedure to that reported for the
synthesis of a nido-carborane analogue of tamoxifen [16].
Deprotection of 8 or 10 with BBr3 were not successful
but treatment of 10 with Co2(CO)8, 1b, afforded 11 as an
oil (Scheme 4). The incorporation of a terminal alkyne
functionality is useful as it can react readily with a number
of organometallic clusters. Reactions of various terminal
alkynes with Os3(CO)10(l-H)2 or 1a [17], as well as with
Ru3(CO)12 to give alkylidyne clusters [18], are known.
The reaction of 1b with alkynes is a well-known reaction
for the protection of the latter in organic syntheses. Fur-
thermore, Co2(CO)6 complexes of terminal alkynes have
been found to be cytotoxic against melanoma and lung car-
cinoma cell lines [19]. Thus, the incorporation of Co2(CO)6
into the tamoxifen skeleton can be expected to impart cyto-
toxic effect against breast cancer cells as well.
The McMurry coupling of disubstituted alkynes has also
been demonstrated previously [20], and we have employed
this in the syntheses outlined in Scheme 5 to afford 12. It
has been shown that 2a can be used to introduce a trios-
mium moiety onto disubstituted alkynes [21], unfortu-
nately, 12 failed to react. On the other hand, 12 reacted
smoothly with 1b to give 13, which was crystallized from
pentane/ether as a deep purple solid. While 13 is purple
in the solid form, its solutions in pentane, ether, chloro-
form, DMSO, and methanol are green. This is unexpected,
as most analogous alkyne–Co2(CO)6 complexes, including
11, are dark red in solution. The solution IR spectrum of
13 displayed three bands as expected of an alkyne–
Co2(CO)6. It is stable in the solid form when kept under

K.H. Chan et al. / Journal of Organometallic Chemistry 691 (2006) 9–19
13
present; the 1H NMR spectrum displayed broad peaks,
suggesting the presence of paramagnetic species. These
observations suggest that Co has been oxidized, resulting
in the release of 12 and loss of CO.
The structure of 13 has been determined by a single crystal
X-ray crystallographic study. The ORTEP diagram showing
the molecular structure, together with selected bond param-
eters, is shown in Fig. 4. Co(1) is symmetrically bound to the
˚
alkyne (1.950(6) and 1.950(5) A for Co(1)–C(1) and Co(1)–
C(2), respectively) but the Co(2) is asymmetrically bound;
the Co(2)–C(2) bond is significantly longer than the
˚
Co(2)—C(1) bond (1.988(5) and 1.959(5) A, respectively.)
˚
The olefinic bond length (C(3)–C(4) = 1.352(8) A) is close
to that for the Co2-coordinated C„C bond length (C(1)–
˚
C(2) = 1.321(8) A), indicating that some multiple bond
character is retained despite binding to Co2(CO)6. As has
been observed for ferrocifen [6e], the two C6H4OH rings are
almost perpendicular to the ethene plane, and the C(421)–
C(4)–C(3), C(411)–C(4)–C(3), and C(2)–C(3)–C(4) bond
angles (122.5(5), 125.3(5) and 128.3(5)ꢁ, respectively) are
larger than 120ꢁ. These have been attributed to the steric
bulk of ferrocene, and in 13 may similarly be attributed
to the steric bulk of the Co2(CO)6 moiety.
˚
Fig. 2. ORTEP diagram and selected bond lengths (A) and angles (ꢁ) for
molecule A of 5h (50% probability thermal ellipsoids.) Os(1A)–Os(2A) =
2.9297(5);
Os(2A)–Os(3A) = 2.8708(5);
Os(1A)–Os(3A) = 2.8562(4);
Os(1A)–O(1A) = 2.146(4); Os(2A)–O(2A) = 2.148(4); O(1A)–C(1A) =
1.253(7); O(2A)–C(1A) = 1.254(7); O(3A)–C(2A) = 1.198(8); C(1A)–
C(2A) = 1.536(8); C(2A)–C(3A) = 1.466(9); O(1A)–Os(1A)–Os(3A) =
91.87(12); O(1A)–Os(1A)–Os(2A) = 80.55(11); Os(3A)–Os(1A)–Os(2A)
=59.477(11);
O(2A)–Os(2A)–Os(3A) = 95.36(12);
O(2A)–Os(2A)–
2.2. Determination of lipophilicity and relative binding
affinity for estrogen receptor
Os(1A) = 80.87(11); Os(3A)–Os(2A)–Os(1A) = 58.987(8); Os(1A)–Os(3A)–
Os(2A) = 61.536(13); O(1A)–C(1A)–O(2A) = 126.8(5).
One of the criteria for a biomimetic is that it has to be
hydrophilic enough to be transported in the bloodstream
to the target tissue, and yet sufficiently hydrophobic to
penetrate the cell membrane. This criterion can be as-
sessed by the determination of the octanol–water partition
coefficient, log(Po/w), using the method of Pomper et al.
[22]; a higher value indicates greater hydrophobicity. The
ideal range for a biomimetic is recognized to be 3.3–5.5;
the reference compound, estradiol, has a value of 3.4.
The relative binding affinity (RBA), is a measure of how
strongly the estrogen receptor recognizes and binds the
tamoxifen mimetic. The biomimetic should have an
RBA as large as possible so that it can compete more
effectively with estradiol for the receptor binding site
R
O
O
Os
n+
Ti
O
H
Os
Os
Fig. 3. Possible chelation leading to loss of carboxylate group.
Ar, but decomposes slowly in solution to form a yellow
solution. This solution showed no carbonyl stretching
bands and reversed phase HPLC indicated that 12 was
and exert its cytotoxic effect. Table 2 shows the log(Po/w
and RBA values determined for some of the tamoxifen
mimetics reported here.
)
Ph
Ph
S
Cl-
Lewis acid
Lewis acid
l
C
O
O
S
-CO
+H +
S
Os
H
Os
Os
Os
H
H
H
Os
Os
Os
Os
Os
7
5l
Scheme 3.

14
K.H. Chan et al. / Journal of Organometallic Chemistry 691 (2006) 9–19
MeO
O
Me
MeO
-
+
Me SiCC Li
3
KOH
OH
OH
THF
MeOH
O
Me Si
3
H
OMe
OMe
OMe
9
8
MeO
MeO
SOCl ,
2
Et N
3
H
H
Co (CO)
2
8
Et O
2
Co (CO)
2
6
H
H
OMe
OMe
11
10
Scheme 4.
The table showed, not unexpectedly, that the simplest
moderate RBA value (RBA = 3.3%). Thus the addition
of the Co2(CO)6 moiety decreased receptor recognition,
which may be attributed to the steric bulk of the cobalt
cluster. Such a diminution in RBA value has also been ob-
served for estradiol on coupling with alkyne–Co2(CO)6
[24]. In comparison with hydroxyferrocifen which has an
RBA of 14.6, it is more poorly recognized by the estrogen
receptor [25]. This could be due to a combination of both
the bulkier and the greater lipophilicity of Co2(CO)6
relative to ferrocifens which may increase the level of
non-specific binding.
triosmium clusters were very hydrophobic. The addition
of hydrocarbon groups to the basic structure (5a) increased
the log(Po/w) very quickly, while addition of polar groups
decreased the log(Po/w) by a relatively smaller value. A
comparison with the values for 1,1-p-dihydroxyphenyl-2-
phenyl-but-1-ene and 1,1-p-dihydroxyphenyl-2-ferrocenyl-
but-1-ene (log(Po/w) = 4.4 and 5.0, respectively) [6f] shows
that while inserting a C„C function or replacing the phe-
nyl ring with a ferrocene group in the former increased the
lipophilicity slightly, introduction of the Co2(CO)6 unit in-
creased it considerably (+1.5 units w.r.t. 12). This suggests
that the carbonyl units were very efficient in increasing the
lipophilicity of the biomimetic.
3. Concluding remarks
Among triosmium cluster compounds, molecules 5d and
5e are closest to diethylstilbestrol. Therefore, these two
compounds should have better affinity toward the receptor
of 17b-estradiol than the others. Unfortunately, 5d and 5e
were not recognized by the receptor (RBA = 0). This result
is more surprising for 5e as this complex has a phenol that
is known to be crucial for the binding to the estrogen recep-
tor [23]. On the other hand, 12, a pure organic compound
exhibited a good affinity for the receptor (RBA = 6.7%)
while 13, its corresponding Co2(CO)6 complex, showed a
We have synthesized a number of simple triosmium clus-
ter derivatives that may mimic tamoxifen. Attempts at
elaboration into the tamoxifen skeleton shows that the var-
ious cluster–ligand linkages used here proved unsuitable
for the McMurry coupling. On the other hand, derivatisa-
tion of the tamoxifen skeleton with an alkyne functionality
prior to introduction of a dicobalt fragment was successful.
Further development of this latter strategy should prove
useful. The lipophilicity and RBA tests conducted on the
compounds synthesized suggest that transition metal car-

K.H. Chan et al. / Journal of Organometallic Chemistry 691 (2006) 9–19
15
OH
OH
PhCC-Li+
THF
McMurry
+
O
O
O
reagent
Cl
Ph
Ph
12
OH
OH
Co2(CO)8
Os3(CO)11(CH3CN)
OH
THF
Et2O
OH
Co2(CO)6
Os3(CO)10
Ph
Ph
13
13'
OH
OH
Scheme 5.
Table 2
logPo/w and relative binding affinity values (RBA) for the estrogen
receptor of selected compounds
Compound
5a
6.2
5b
7.1
5c
6.6
5d
5e
5f
5j
12
13
log(Po/w
RBAa
a
)
7.7
0.0
7.0
0.0
8.3
7.0
5.0
6.7
6.5
3.3
As % for ERa in DMSO, 4 ꢁC, 3 h.
hydrophilic groups, and possibly placing the organometal-
lic fragment further from the tamoxifen skeleton.
4. Experimental
All manipulations were carried out using standard
Schlenk techniques under an argon or nitrogen atmosphere
[24]. Solvents that were used for reactions were distilled over
the appropriate drying agents under nitrogen before use.
Reaction mixtures were separated by flash column chroma-
tography or TLC using silica gel (impregnated with fluores-
cent indicator; 254 nm) pre-coated on glass plates (0.25 mm
layer thickness). The clusters 2a, 2b and 3 were prepared
according to reported procedures [25]. All other reactants
and reagents were purchased and used as supplied without
further purification. 1D and 2D NMR spectra were recorded
on Bruker ACF-300 and Bruker AMX-500 FT-NMR spec-
trometers, respectively. All spectra were recorded as CDCl3
solutions at 300 K unless otherwise stated. Mass spectromet-
ric determinations were performed on a VG Micromass 7035
˚
Fig. 4. ORTEP diagram and selected bond lengths (A) and angles (ꢁ)
for 13 (50% probability thermal ellipsoids.) Co(1)–Co(2) = 2.4510(11);
Co(1)–C(1) = 1.950(5); Co(1)–C(2) = 1.950(6); Co(2)–C(1) = 1.959(5);
Co(2)–C(2) = 1.988(5); Co(1)–C(11) = 1.822(7); Co(1)–C(12) = 1.813(7);
Co(1)–C(13) = 1.788(7); Co(2)–C(21) = 1.813(7); Co(2)–C(22) = 1.822(7);
Co(2)–C(23) = 1.796(7); C(1)–C(2) = 1.321(8); C(2)–C(3) = 1.475(8);
C(3)–C(4) = 1.352(8); C(2)–C(3)–C(5) = 115.8(5); C(421)–C(4)–C(411) =
112.2(5).
bonyl groups increase the lipophilicity dramatically and
also reduce RBA via steric hindrance. This would mean
that potential molecular targets should incorporate more

16
K.H. Chan et al. / Journal of Organometallic Chemistry 691 (2006) 9–19
mass spectrometer coupled with a Hewlett–Packard 5890A
GC system or Nermag R 10–10C. Microanalyses were car-
ried out by the elemental analysis laboratory at the Depart-
ment of Chemistry, National University of Singapore.
Lipophilicity constants were determined on a System Gold
HPLC System (Beckman Coulter, Inc) equipped with a
Photo Diode Array detector (277 nm) and a 32 Karat Work-
station (version 3.0).
was allowed to stir at the same temperature for 15 min. A
solution of 2-(4-methoxyphenyl)-40-methoxyacetophenone
(1.026 g, 3.92 mmol) in THF (50 mL) was then injected
dropwise into the freshly generated (trimethylsilyl)acetylide
lithium reagent maintained at ꢀ78 ꢁC, after which the mix-
ture was allowed to warm up to ambient temperature with
stirring. After 14 h, the reaction was quenched with satu-
rated aqueous ammonium chloride (5 mL). The aqueous
fraction was extracted with diethylether (3 · 20 mL). The
combined ethereal fraction was washed with saturated
aqueous brine (3 · 20 mL) and dried with magnesium sul-
fate for 30 min. After the solvent was removed in vacuo,
the colourless oil was purified by flash column chromatog-
raphy to yield a white solid.
Yield: 1.040 g (73.2%). Rf: 0.49 (ether/pentane; 1/5 v/v).
1H NMR: d 7.49 (d, 2H, C6H4, 3J = 8.92 Hz), 7.08 (d, 2H,
C6H4, 3J = 8.69 Hz), 6.87 (d, 2H, C6H4), 6.79 (d, 2H,
C6H4), 3.82 (s, 3H, OCH3), 3.79 (s, 3H, OCH3), 3.06 (d,
2H, CH2, 4J = 2.27 Hz), 2.37 (s, 1H, OH), 0.19 (s, 9H,
Si(CH3)3). 13C NMR: d 159.3, 158.8, 136.2, 132.2, 128.1,
127.1, 113.5, 113.3, 107.8, 91.9 (aromatic); 73.5 (COH);
55.5, 55.4 (MeO); 51.1 (CH2); 0.0 (Me3Si). EI-MS (m/z,
rel. abundance): 354.1 (<1, M+), 336.1 (60.0, M+ ꢀ H2O).
HR-EI: Calc. for C21H26O3Si: 354.1651. Found: 354.1642.
4.1. General procedure for the syntheses of 5
To a solution of 3 and the carboxylic acid (excess) in
dichloromethane (0.5 mL) was added a drop of HBF4 Æ
Et2O, resulting in a white precipitate that may disappear
on prolonged stirring. The mixture was allowed to stir at
ambient temperature for 30 min before purification by
thin-layer chromatography.
4.2. Alternative syntheses of 5e and 5k
To 2b (200 mg, 0.214 mmol) in THF (20 mL) was added
4-hydroxycinnamic acid (40.3 mg, 0.245 mmol), after
which the Carius tube was sealed and evacuated with three
cycles of freeze–pump–thaw. The mixture was then heated
at 60 ꢁC for 3 h. After the solvent was removed in vacuo,
the residue was re-dissolved in 0.5 mL of dichloromethane
and separated by TLC. Yield of 5e = 30.7 mg (14.1%).
A similar procedure was followed for the synthesis of 5k
from 2b (200 mg, 0.214 mmol) and 4-acetoxybenzoic acid
(50.3 mg, 0.307 mmol). Yield = 11.5 mg (5.3%).
4.5. Synthesis of 9
To 8 (2.35 g, 6.64 mmol) in methanol (25 mL) was
added potassium hydroxide powder (0.451 g, 8.03 mmol)
at ambient temperature, after which the mixture was al-
lowed to stir for 3 h. After methanol was removed in va-
cuo, the colourless oil was partitioned between a mixture
of diethylether (20 mL) and water (20 mL). The aqueous
fraction was extracted with diethylether (3 · 20 mL). The
combined ethereal fraction was washed with saturated
aqueous brine (3 · 20 mL) and dried with magnesium sul-
fate for 30 min. After the solvent was removed in vacuo,
the white residue was purified by flash column chromatog-
raphy to yield a white solid. Yield: 1.596 g (85.2%). Rf: 0.27
4.3. General procedure for McMurry reactions of 5
To a suspension of Mg/Zn powder in THF (10 ml) was in-
jected TiCl4 dropwise at 0 ꢁC, after which the mixture was re-
fluxed for 2 h to a black suspension. An equimolar solution
of 5 and 4,40-dihydroxybenzophenone in THF (20 ml) was
then injected into the freshly generated McMurry reagent
at 0 ꢁC. After refluxing the mixture for 5 h, the reaction
was quenched with saturated aqueous sodium carbonate
(100 mL). The aqueous fraction was extracted with ethyl
acetate (3 · 50 mL). The combined yellowish ethereal frac-
tion was washed with saturated aqueous brine (3 · 30 mL)
and dried with sodium sulfate for 30 min. After the solvent
was removed in vacuo, the residue was either re-dissolved
1
(ether/pentane; 1/4 v/v). H NMR: d 7.50 (d, 2H, C6H4,
3
3J = 8.96 Hz), 7.08 (d, 2H, C6H4, J = 8.75 Hz), 6.87 (d,
2H, C6H4), 6.80 (d, 2H, C6H4), 3.82 (s, 3H, OCH3), 3.79
(s, 3H, OCH3), 3.11 (s, 2H, CH2), 2.67 (s, 1Hh, C„CAH),
2.41 (s, 1H, OH). 13C NMR: d 159.2, 158.7, 135.8, 131.9,
127.6, 126.8, 113.4, 113.3; 86.1 (CAC„); 75.1 (C„CH);
72.8 (CAC„); 55.3, 55.1 (MeO); 50.6 (CH2). EI-MS
(m/z, rel. abundance, ion): 282.1 (19, M+), 264.1 (59,
M+ ꢀ H2O), 249.1 (47, M+ ꢀ H2O ꢀ Me). HR-EI: Calc.
for C18H18O3: 282.1256. Found: 282.1255.
1
in CDCl3 and analyzed by H NMR (5c, 5k), or separated
by TLC (5g, 5h, 5i, 5j, 5l). For 5g–5j, cluster 6 was isolated as
the main product in their respective yields: 5g (11.6 mg,
56.1%), 5h (9.8 mg, 47.1%), 5i (44.0 mg, 67.9%), and 5j
(35.0 mg, 52.6%). For 5l, cluster 7 was isolated as the main
product (24.1 mg, 69.6%).
4.6. Synthesis of 10
4.4. Synthesis of 8
To 9 (0.658 g, 2.33 mmol) in triethylamine (6 mL) and
diethylether (2 mL) was injected dropwise thionyl chloride
(0.424 g, 3.56 mmol) [11] at 0 ꢁC. A yellowish-brown solid
precipitated gradually and the suspension darkened over
time. After stirring at ambient temperature for 2 h, the
To (trimethylsilyl)acetylene (0.588 g, 6.00 mmol) in THF
(10 mL) was injected n-butyllithium (4.2 mL, 1.6 M,
6.72 mmol) dropwise at ꢀ78 ꢁC, after which the solution

K.H. Chan et al. / Journal of Organometallic Chemistry 691 (2006) 9–19
17
reaction was quenched with water (20 mL). The aqueous
fraction was extracted with diethylether (3 · 10 mL). The
combined ethereal fraction was washed with saturated
aqueous brine (3 · 20 mL) and dried with magnesium sul-
fate for 30 min. After the solvent was removed in vacuo,
the dark red oil was purified by flash column chromatogra-
phy to yield a white solid. Yield: 0.488 g (79.2%). Rf: 0.50
CH3CH2).13C NMR: d 188.5 (CO); 133.0, 130.6, 128.6,
120.0 (aromatic); 90.6 („CCO); 87.6 (C„CCO); 38.8
(CH2); 8.1 (CH3).
4.9. Synthesis of 12
To a suspension of Zn powder (25.5 mg, 0.392 mmol) in
THF (20 mL) was injected TiCl4 (35 mg, 0.182 mmol)
dropwise at 0 ꢁC, after which the mixture was refluxed
for 2 h to a black suspension. A solution of 5c (24.0 mg,
25.6 lmol) in THF (10 mL) was then injected into the
freshly generated McMurry reagent at 0 ꢁC. After refluxing
the mixture for 2 h, the reaction was quenched with satu-
rated aqueous sodium carbonate (100 mL). The aqueous
fraction was extracted with ethyl acetate (3 · 50 mL). The
combined yellowish ethereal fraction was washed with sat-
urated aqueous brine (3 · 30 mL) and dried with sodium
sulfate for 30 min. After the solvent was removed in vacuo,
the yellow oil was purified by TLC. Yield: 1.987 g (71.7%).
1
(ether/pentane; 1/5 v/v). H NMR: d 7.94 (d, 2H, C6H4,
3
3J = 8.70 Hz), 7.63 (d, 2H, C6H4, J = 9.00 Hz), 7.10 (s,
1H, C@CAH), 6.91 (d, 4H, C6H4), 3.85 (s, 3H, OCH3),
3.84 (s, 3H, OCH3), 3.52 (s, 1H, C„CAH). 13C NMR: d
159.6, 159.3, 130.4, 129.3, 127.4, 117.7, 113.8, 113.7 (aro-
matic); 132.0 (C@CH); 134.3 (@CH); 84.7 (HC„); 82.9
(CAC„); 55.4, 55.3 (MeO). EI-MS (m/z, rel. abundance,
ion): 264.1 (100, M+), 249.1 (91, M+ ꢀ Me). HR-EI: Calc.
for C18H16O2: 264.1150. Found: 264.1152.
4.7. Synthesis of 11
1
To 10 (0.207 g, 0.783 mmol) in diethylether (5 mL) was
added dicobalt octacarbonyl (0.268 g, 0.783 mmol), result-
ing in the colour of the solution to change from colourless
to dark red immediately. The mixture was then allowed to
stir at ambient temperature for 30 min. After the reaction,
the suspension was filtered through a sintered glass frit and
the dark red filtrate was concentrated in vacuo to a viscous
solution. Pentane (1 mL) was added to dilute the solution,
after which it was sealed under argon and cooled to 4 ꢁC.
11 was isolated as a dark red oil. IR (ether, mCO, cmꢀ1):
Rf: 0.42 (ether/pentane; 2/1 v/v). H NMR (CD3OD): d
7.27 (d, 2H, C6H4, 3J = 8.67 Hz), 7.27–7.23 (m, 5H,
C6H5), 6.95 (d, 2H, C6H4, 3J = 8.58 Hz), 6.77 (d, 2H,
C6H4), 6.72 (d, 2H, C6H4), 2.32 (q, 2Hb,CH2CH3,
3J = 7.46 Hz), 1.20 (t, 3H, CH2CH3). 13C NMR (CD3OD):
d 157.9, 148.7, 135.4, 134.2, 132.6, 132.1, 131.9, 129.3,
125.5, 115.8, 115.1; 128.7 (EtC@C); 121.4 (EtC@); 93.2
(PhC„C); 92.8 (PhC„C); 28.7 (CH2); 14.2 (CH3). Calc.
for C24H20O2 (%): C, 84.68; H, 5.92. Found: C, 84.31; H,
5.87.
1
2092 (s), 2051 (vs), 2028 (vs). H NMR: d 7.34 (d, 2H,
C6H4, 3J = 7.16 Hz), 7.24 (d, 2H, C6H4, 3J = 7.35 Hz), 6.93
(d, 2H, C6H4), 6.89 (d, 2H, C6H4), 6.81 (s, 1H, C@CAH),
5.95 (s, 1H, C„CAH), 3.85 (s, 3H, OCH3), 3.82 (s, 3H,
OCH3). EI-MS (m/z): 522 (M+ ꢀ CO), 466 (M+ ꢀ 3CO),
438 (M+ ꢀ 4CO), 410 (M+ ꢀ 5CO), 382 (M+ ꢀ 6CO), 323
(M+ ꢀ 6CO ꢀ Co), 264 (M+ ꢀ 6CO ꢀ 2Co).
4.10. Synthesis of 13
To 12 (0.338 g, 0.993 mmol) in diethylether (5 mL) was
added Co2(CO)8 (0.335 g, 0.993 mmol), resulting in the col-
our of the solution to change from colourless to dark green
immediately. The mixture was then allowed to stir at ambi-
ent temperature for 30 min. After the reaction, the suspen-
sion was filtered through a sintered glass frit and the dark
green filtrate was concentrated in vacuo to a viscous solu-
tion. Pentane (1 mL) was added to dilute the solution, after
which it was sealed under argon and recrystallized in the
fridge. Yield: 0.366 g (58.8%). IR (ether, mCO, cmꢀ1): 2085
4.8. Synthesis of 1-phenyl-1-pentyn-3-one
To phenylacetylene (2.131 g, 20.9 mmol) in THF
(30 mL) was injected n-butyllithium (14.3 mL, 23.0 mmol;
1.6 M) dropwise at ꢀ78 ꢁC, after which the solution was al-
lowed to warm up to room temperature with stirring. The
freshly generated phenylacetylide lithium reagent was then
injected dropwise into a solution of propionyl chloride
(7.736 g, 83.6 mmol) in THF (30 mL) at ꢀ78 ꢁC. After stir-
ring at ꢀ78 ꢁC for 2 h, the reaction was quenched slowly
with water (20 mL). The aqueous fraction was extracted
with diethylether (3 · 20 mL). The combined ethereal frac-
tion was then washed with saturated aqueous brine
(3 · 20 mL) and dried with magnesium sulfate for 30 min.
After the solvent was removed in vacuo, the light yellow
oil was purified by flash column chromatography to yield
a colourless liquid. Yield: 3.023 g (91.4%). Rf: 0.57 (ether/
pentane; 1/5 v/v). 1H NMR: d 7.57 (dd, 2H, o-C6H5,
3J = 8.14 Hz, 4J = 1.57 Hz), 7.46–7.35 (m, 3H, m,p-
1
(m), 2048 (vs), 2024 (s), 2015 (sh, m). H NMR (CD3OD):
d 7.13–7.05 (m, 3H, m,p-C6H5), 6.96 (d, 4H, 3J = 7.32 Hz),
6.73 (d, 2H, 3J = 8.63 Hz), 6.72 (d, 2H, 3J = 8.63 Hz), 6.27
(d, 2H, 3J = 8.62 Hz), 2.58 (q, 2H, CH2CH3, 3J = 7.45 Hz),
0.95 (t, 3H, CH2CH3). Calc. for C30H20O8Co2 (%): C,
57.53; H, 3.22. Found: C, 57.78; H, 3.83.
5. Crystal structure determinations
Crystals were grown from dichloromethane/hexane
solutions and mounted on quartz fibres. X-ray data were
collected on a Bruker AXS APEX system, using Mo Ka
radiation, at 223 K with the SMART suite of programs
[26]. Data were processed and corrected for Lorentz and
polarization effects with SAINT [27], and for absorption
3
C6H5), 2.70 (q, 2H, CH3CH2, J = 7.35 Hz), 1.22 (t, 3H,

18
K.H. Chan et al. / Journal of Organometallic Chemistry 691 (2006) 9–19
effects with SADABS [28]. Structural solution and refinement
were carried out with the SHELXTL suite of programs [29].
Crystallographic and refinement data are given in Table 3.
The structures were solved by either direct methods or
Patterson maps to locate the heavy atoms, followed by dif-
ference maps for the light, non-hydrogen atoms. There
were two crystallographically independent molecules each
in the asymmetric units for 6 and 13. There was a partial
water molecule in 13, which was modeled as disordered
over three sites of equal occupancies. All non-hydrogen
atoms were generally given anisotropic displacement
parameters in the final model, except for the partial water
molecule in 13, for which the O atoms were given equiva-
lent isotropic thermal parameters.
6.3. Determination of the relative binding affinity of the
compounds for the estrogen receptor
RBA values were measured on ERa from lamb uterine
cytosol. Sheep uterine cytosol were prepared in buffer A
(0.05 M Tris–HCl, 0.25 M sucrose, 0.1% b-mercap-
toethanol, pH 7.4 at 25 ꢁC) as described earlier [30]. Ali-
quots (200 ll) of cytosol were incubated for 3 h at 0 ꢁC
with [6,7-3H] estradiol (2 · 10ꢀ9 M, specific activity
1.62 TBq/mmol, NEN Life Science, Boston, MA) in the
presence of nine concentrations of the hormones to be
tested. At the end of the incubation period, the free and
bound fractions of the tracer were separated by protamine
sulfate precipitation. The percentage reduction in binding
of [3H] estradiol (Y) was calculated using the logit transfor-
mation of Y (logitY: ln[y/1 ꢀ Y]) vs. the log of the mass of
the competing steroid. The concentration of unlabeled ste-
roid required to displace 50% of the bound [3H] estradiol
was calculated for each steroid tested, and the results ex-
pressed as RBA. The RBA value of estradiol is by defini-
tion equal to 100%.
6. Biochemical experiments
6.1. Materials
17b-Estradiol and OH-Tam (Z + E) were obtained from
Sigma–Aldrich (France). Stock solutions (1 · 10ꢀ3 M) of
the compounds to be tested were prepared in DMSO and
were kept at ꢀ20ꢁC in the dark.
6.4. Measurement of octanol/water partition coefficient
(logPo/w) of the compounds
6.2. Animal tissues
The logPo/w values of the compounds were determined by
reversed-phase HPLC on a C-8 column (nucleosil 5.C8, from
Macherey Nagel, France) according to the method previ-
ously described by Pomper [22]. Measurement of the chro-
matographic capacity factors (k0) for each compounds was
Sheep uteri weighing approximately 7 g were obtained
from the slaughterhouse at Mantes-la-Jolie, France. They
were immediately frozen and kept in liquid nitrogen prior
to use.
Table 3
Crystal and refinement data for 5h, 6 and 13
Compound
5h
C
1000.83
6
13
C30H20Co2O8.0.19 H2O
629.70
Tetragonal
P4/ncc
25.6996(4)
25.6996(4)
19.4047(6)
90
Empirical formula
Formula weight
Crystal system
Space group
18H6O13Os3
C10HClO10Os3
887.16
Triclinic
P1
9.1060(12)
14.0244(18)
14.2271(18)
70.243(2)
89.376(3)
76.668(2)
Triclinic
ꢀ
ꢀ
P1
˚
a (A)
9.6586(14)
12.8849(18)
18.699(3)
83.516(3)
78.168(3)
˚
b (A)
˚
c (A)
a (ꢁ)
b (ꢁ)
c (ꢁ)
90
90
86.055(3)
2260.6(6)
3
˚
Volume (A )
1659.4(4)
12816.2(5)
Z
4
4
16
1.305
1.079
5118
0.23 · 0.06 · 0.03
2.06–26.37
180992
Density (calculated) (Mg/m3)
Absorption coefficient (mmꢀ1
F(000)
2.941
16.886
1784
0.34 · 0.28 · 0.10
2.04–30.01
35247
3.551
23.121
1544
0.28 · 0.18 · 0.08
2.30–30.04
25368
)
Crystal size (mm3)
H range for data collection (ꢁ)
Reflections collected
Independent reflections [R(int)]
Max. and min. transmission
Data/restraints/parameters
Goodness-of-fit on F2
13113 [0.0395]
0.2830 and 0.0691
13113/0/613
1.058
9424 [0.0444]
0.2592 and 0.0595
9424/0/439
1.026
6564 [0.1383]
0.9683 and 0.7894
6564/0/368
1.338
Final R indices [I > 2r(I)]
R indices (all data)
Largest diff. peak and hole (e A
R1 = 0.0318, wR2 = 0.0635
R1 = 0.0406, wR2 = 0.0664
2.005 and ꢀ1.793
R1 = 0.0315, wR2 = 0.0670
R1 = 0.0418, wR2 = 0.0701
1.654 and ꢀ1.854
R1 = 0.1011, wR2 = 0.2419
R1 = 0.1191, wR2 = 0.2534
0.973 and ꢀ0.697
ꢀ3
˚
)

K.H. Chan et al. / Journal of Organometallic Chemistry 691 (2006) 9–19
19
(b) S. Top, B. Dauer, J. Vaissermann, G. Jaouen, J. Organomet.
Chem. 541 (1997) 355;
done at various concentrations in the range 85–60% metha-
nol (containing 0.25% octanol) and an aqueous phase
consisting of 0.15% n-decylamine in 0.02 M MOPS (3-
morpholinopropanesulfonic acid) buffer pH 7.4 (prepared
in 1-octanol–saturated water). These capacity factors (k0)
were extrapolated to 100% of the aqueous component given
the value of k0w. logPo/w (y) was then obtained by the for-
mula: log Po=w ¼ 0:13418 þ 0:98452 log k0w.
`
(c) G. Jaouen, S. Top, A. Vessieres, G. Leclerq, J. Quivy, L. Jin, A.
Croisy, C. R. Acad. Sci. Paris Ser. IIc (2000) 89;
`
(d) S. Top, A. Vessieres, C. Cabestaing, I. La¨ıos, G. Leclerq, C.
Provot, G. Jaouen, J. Organomet. Chem. 637 (2001) 500;
`
(e) S. Top, A. Vessieres, G. Leclercq, J. Quivy, J. Tang, J.
Vaissermann, M. Huche´, G. Jaouen, Chem. Eur. J. 9 (2003) 5223;
`
(f) A. Vessieres, S. Top, P. Pigeon, E. Hillard, L. Boubekeur, D.
Spera, G. Jaouen, J. Med. Chem. 48 (2005) 3937.
`
´
[7] P. Pigeon, S. Top, A. Vessieres, M. Huche, E.A. Hillard, E. Salomon,
Acknowledgments
G. Jaouen, J. Med. Chem. 48 (2005) 2814.
[8] D. Roberto, E. Lucenti, C. Roveda, R. Ugo, Organometallics 16
(1997) 5974.
[9] E.W. Ainscough, A.M. Brodie, R.K. Coll, B.A. Coombridge, J.M.
Waters, J. Organomet. Chem. 556 (1998) 197.
[10] A.D. Humphries, H.D. Kaesz, Prog. Inorg. Chem. 2 (1979) 145.
[11] (a) M.W. Lum, W.K. Leong, J. Chem. Soc., Dalton Trans. (2001)
2476;
This work was supported by the National University of
Singapore (Research Grant No. R143-000-190-112),
CNRS and MRT (France). One of us (K.H.C.) thanks
the Kiang Ai Kim Foundation for a Research Scholarship.
(b) M.W. Lum, W.K. Leong, Inorg. Chim. Acta 357 (2004) 769.
[12] A.J. Deeming, S. Hasso, J. Organomet. Chem. 114 (1976) 313.
[13] A.A. Colletta, J.R. Benson, M. Baum, Breast Cancer Res. Treat. 31
(1994) 5.
[14] P. Finocchiaro, E. Libertini, A. Recca, La Chim. E LꢀIndus 64 (1982)
644.
[15] E.W. Ainscough, A.M. Brodie, R.K. Coll, B.A. Coombridge, J.M.
Waters, J. Organomet. Chem. 556 (1998) 197.
[16] J. Valliant, P. Schaffer, K.A. Stephenson, J.F. Britten, J. Org. Chem.
67 (2002) 383.
Appendix A. Supplementary data
Crystallographic data (excluding structure factors) for the
structures in this paper have been deposited with the Cam-
bridge Crystallographic Data Centre as supplementary publi-
cation numbers CCDC 280441–280443. Copies of the data
can be obtained, free of charge, on application to CCDC, 12
Union Road, Cambridge CB2 1EZ, UK, (fax: +44 1223
336033 or e-mail: deposit@ccdc.cam.ac.uk). Supplementary
data associatedwith this article can be found, in the online ver-
sion, at doi:10.1016/j.jorganchem.2005.08.041.
[17] (a) A.J. Deeming, S. Hasso, M. Underhill, J. Organomet. Chem. 80
(1974) C53;
(b) O. Gambino, R.P. Ferrari, M. Chinone, G.A. Vaglio, Inorg.
Chim. Acta 12 (1975) 155.
[18] (a) E. Sappa, O. Gambino, L. Milone, G. Cetini, J. Organomet.
Chem. 39 (1972) 169;
References
(b) A.A. Koridze, A.I. Yanovsky, Y.T. Struchkov, J. Organomet.
Chem. 441 (1992) 277, and references therein.
[19] M. Jung, D.E. Kerr, P.D. Senter, Arch. Pharm. Pharm. Med. Chem.
330 (1997) 173.
[20] M. Periasamy, G. Srinivas, G.V. Karunakar, P. Bharathi, Tetrahe-
dron Lett. 40 (1999) 7577.
[21] R.D. Adams, O.-S. Kwon, B. Qu, M.D. Smith, Organometallics 20
(2001) 5225.
[1] (a) V.C. Jordan, Pharmacol. Rev. 36 (1984) 245;
(b) H.Y.P. Lam, Biochem. Biophys. Res. Commun. 118 (1984) 27;
(c) C.A. OꢀBrian, R.M. Liskamp, D.H. Solomon, I.B. Weinstein,
Cancer Res. 45 (1985) 2462;
(d) L.J. Brandes, R.P. Bogdanovic, M.D. Cawker, R. Bose, Cancer
Chemother. Pharmacol. 18 (1986) 21;
(e) V.C. Jordan, J. Med. Chem. 46 (2003) 883;
(f) V.C. Jordan, J. Med. Chem. 46 (2003) 1081.
[22] M.G. Pomper, H. van Brocklin, M.M. Thieme, R.D. Thomas, D.O.
Kiesewetter, K.E. Carlson, C.J. Mathias, M.J. Welsh, J.A. Katzen-
ellenbogen, J. Med. Chem. 33 (1990) 3143.
[23] J. Raus, H. Martens, G. Leclercq, in: Cytotoxic Estrogens in
Hormone Receptive Tumors, Academic, London, 1980.
[24] D. Shriver, M.A. Drzedon, The Manipulation of Air-sensitive
Compounds, second ed., Wiley, New York, 1996.
[25] (a) B.F.G. Johnson, J. Lewis, Inorg. Synth. 28 (1989) 92;
(b) D. Roberto, E. Lucenti, C. Roveda, R. Ugo, Organometallics 16
(1997) 5974.
`
[2] S. Top, E.B. Kaloun, A. Vessieres, G. Leclerq, I. La¨ıos, M. Ourevitch,
C. Deuschel, M.J. McGlinchey, G. Jaouen, ChemBioChem 4 (2003)
754.
[3] (a) Y. Endo, T. Iijima, Y. Yamakoshi, M. Yamaguchi, H. Fukusawa,
K. Shudo, J. Med. Chem. 42 (1999) 1501;
(b) M.F. Hawthorne, M.T. Groudine, US Patent 6,074,625, 2000;
(c) J.F. Valliant, P. Schaffer, K.A. Stephenson, J.F. Britten, J. Org.
Chem. 67 (2002) 383.
`
´
[4] (a) S. Top, A. Vessieres, P. Pigeon, M.-N. Rager, M. Huche, E. Salomon,
C. Cabestaing, J. Vaissermann, G. Jaouen, ChemBioChem 5 (2004) 1104;
[26] SMART version 5.628, Bruker AXS Inc., Madison, WI, USA, 2001.
[27] SAINT+ version 6.22a, Bruker AXS Inc., Madison, WI, USA, 2001.
[28] G.M. Sheldrick, SADABS, 1996.
`
(b) G. Jaouen, S. Top, A. Vessieres, P. Pigeon, G. Leclerq, I. La¨ıos, J.
Chem. Soc., Chem. Commun. (2001) 383.
`
[5] S. Top, E.B. Kaloun, A. Vessieres, I. La¨ıos, G. Leclerq, G. Jaouen,
[29] SHELXTL version 5.03, Siemens Energy & Automation Inc., Madison,
WI, USA.
J. Organomet. Chem. 643–644 (2002) 350.
`
`
[6] (a) S. Top, J. Tang, A. Vessieres, D. Carrez, C. Provot, G. Jaouen, J.
[30] A. Vessieres, S. Top, A.A. Ismail, I.S. Butler, M. Louer, G. Jaouen,
¨
Chem. Soc., Chem. Commun. (1996) 955;
Biochemistry 27 (1988) 6659.