Water soluble, hydrolytically stable derivatives of the antitumor drug titanocene dichloride and binding studies with nucleotides

Article abstract of DOI:10.1016/S0022-328X(98)00441-0

The rate of hydrolysis of the aromatic rings of Cp₂TiX₂ and the dimethylsubstituted derivatives (MeCp)₂TiX₂ [X=Cl, O₂CCH₂NH₃Cl], in aqueous solutions at pH 2-8 have been studied

Full text of DOI:10.1016/S0022-328X(98)00441-0

Journal of Organometallic Chemistry 565 (1998) 29–35
Water soluble, hydrolytically stable derivatives of the antitumor drug
titanocene dichloride and binding studies with nucleotides1
George Mokdsi, Margaret M. Harding *
School of Chemistry, Uni6ersity of Sydney, Sydney, N.S.W. 2006, Australia
Received 8 November 1997
Abstract
The rate of hydrolysis of the aromatic rings of Cp2TiX2 and the dimethylsubstituted derivatives (MeCp)2TiX2 [X=Cl,
O2CCH2NH3Cl], in aqueous solutions at pH 2–8 have been studied by 1H-NMR spectroscopy. Rapid hydrolysis of both the
halide and cyclopentadienyl ligands in Cp2TiX2 [X=Cl, O2CCH2NH3Cl] occurs to give predominantly insoluble precipitates at
pH 7. In contrast, under the same experimental conditions, the predominant species present in aqueous solutions of (MeCp)2TiX2
[X=Cl, O2CCH2NH3Cl] at pH 2–8 contains both methycyclopentadienyl rings metal bound. At pHB5, Cp2TiX2 and
(MeCp)2TiX2 form similar complex(es) with purine nucleotides. However, at pH\5, while stable adducts between nucleotides and
Cp2TiX2 are not formed, in the presence of 1 equiv. of 5%-dAMP or 5%-dGMP, (MeCp)2TiX2 formed complex(es) which were stable
for 24 h. These results suggest that formation of stable chelates between (MeCp)2TiX2 and nucleic acid constituents in vivo is
possible. © 1998 Elsevier Science S.A. All rights reserved.
Keywords: Titanocene dichloride; Bis(methylcyclopentadienyl) titanocene dichloride; Antitumor; Nucleotide-binding; Cp hydroly-
sis; NMR spectroscopy
1. Introduction
species have been largely unsuccessful due to both the
poor solubility of 1 in water and the hydrolysis of the
Cp rings, a reaction which is accelerated at high pH
and results in precipitation of uncharacterised hydroly-
sis products [8]. Samples have been administered in 9:1
saline:DMSO solutions at low pH [9]; preparation of
aqueous solutions at pH\4.5 is not possible as increas-
ing the pH is accompanied by formation of insoluble,
uncharacterised precipitates [7,8].
Prompted by initial reports that administration of
buffered solutions of titanocene dichloride 1 reduced
the side-effects associated with treatment [10], and in an
effort to identify the biologically active species, we set
out to prepare titanocene derivatives which were rela-
tively stable at pH 7.0 and demonstrated appreciable
solubility in water. This paper reports the synthesis,
studies of the solubility and hydrolytic stability, and
binding to nucleotides of bis(methylcyclopentadienyl)
Titanocene dichloride (Cp2TiCl2; Cp=p5-C5H5) 1
has been shown to exhibit antitumor properties against
a range of human and murine tumours [1]. The drug
entered phase I clinical trials in late 1991 with initial
results showing a lack of cross-reactivity against plat-
inum base drugs [2]. The results of further clinical trials
are required in order to assess whether the drug has
potential as a viable anticancer drug.
The mechanism of antitumour action of titanocene
dichloride 1 has not been determined, but interaction
with DNA has been implicated in the mechanism of
action [3–7]. Efforts to identify the biologically active
* Corresponding author.
1 Dedicated to Professor Michael Bruce on the occasion of his 60th
birthday.
0022-328X/98/$19.00 © 1998 Elsevier Science S.A. All rights reserved.
PII S0022-328X(98)00441-0

30
G. Mokdsi, M.M. Harding / Journal of Organometallic Chemistry 565 (1998) 29–35
titanocene dichloride (MeCp)2TiCl2 3 and the bisglycine
analogue (MeCp)2Ti(O2CCH2NH3Cl)2 4 in comparison
to the unsubstituted metallocenes Cp2TiCl2 1 and
Cp2Ti(O2CCH2NH3Cl)2 2, respectively. Alkyl substitu-
tion of the Cp rings has been reported to improve the
hydrolytic stability of titanocene dichloride [11], but
introduction of hydrophobic groups generally decreases
aqueous solubility. Introduction of glycine ligands [12]
gives the highly water soluble complex 4 in which the Cp
rings remain metal bound at physiological pH and which
forms stable complexes with nucleotides at pH 7.
ligands of metallocenes with charged or hydrophilic
ligands in order to improve aqueous solubility, including
the synthesis and improved solubility of complex 2, has
been reported previously [12,14].
2.3. Hydrolysis studies
The rate of halide and Cp ring hydrolysis in titanocene
dichloride 1 has been well-characterised [8]. Rapid hy-
drolysis of the two halide ligands occurs in a series of
equilibrium reactions to give a solution at pH ca. 2 in
which the Cp rings remain metal bound for \24 h.
However, at higher pH values protonolysis of the Cp
rings occurs to give cyclopentadiene and dicyclopentadi-
ene as well as insoluble, uncharacterised hydrolysis
products. The exact nature of the pseuodohalide ligands
have not been identified with Cp2TiCl2(aq) likely to
2. Results and discussion
2.1. Synthesis of complexes
Bis(methylcyclopentadienyl) titanocene dichloride 3
was prepared using a similar procedure to that reported
in the literature [13]. Thus, methylcyclopentadiene, gen-
erated from freshly cracked dimethylcyclopentadiene,
was treated with sodium sand followed by titanium(IV)
tetrachloride to give 3 in 25% yield. As reported, consid-
erable difficulty was experienced in obtaining 3 in high
purity due to the presence of significant amounts of
dicylopentadiene in commercially available methylcy-
clopentadiene dimer [12]. Careful, multiple distillations
were the most effective way of obtaining highly pure
methylcyclopentadiene and avoid formation of minor
amounts of Cp(MeCp)TiCl2 along with 3 in the reaction.
The glycine analogue 4 was prepared in 75% yield
using a similar procedure to that reported for the
preparation of titanocene bis(glycine) 2 [12]. Thus, treat-
ment of bis(methylcyclopentadienyl) titanocene dichlo-
ride 3 with glycine (2 eq.) in methanol at r.t. afforded the
orange complex 4. The complex was washed with chlo-
roform to remove any unreacted starting material to give
4 which was pure by microanalysis and NMR spec-
troscopy. In our hands the success of the reaction was
highly variable and frequently material was isolated that
co-crystallised with unreacted glycine. Attempts to fur-
ther purify such material was generally unsuccessful.
contain a number of species such as Cp2Ti(H2O)Cl+, or
+
Cp2Ti(H2O)x(OH)y(2−y)
.
While the synthesis of 2 has been reported [13],
complete hydrolysis studies in water have not been
carried out. In the solid state the complex crystallises
from water in an oligomeric form in which the glycine
carboxylates are coordinated to the metal [12]. This fact,
taken with the downfield chemical shift of the glycine
1
methylene in the H-NMR spectrum of 2 (3.38 ppm)
compared with unbound glycine (3.68 ppm) was taken
as evidence that the glycine ligands are directly coordi-
nated to the metal centre in solution as well as the solid
state [12]. The reported changes in chemical shift on
coordination of the glycine carboxylate to the metal
centre do not appear to have taken into account the
change in solution pH. We have carried out careful
comparison of the NMR spectra of free glycine at pD
2–8 with a solution of Cp2Ti(O2CCH2NH3Cl)2 at the
same pD values. These experiments showed that there is
no change in chemical shift of free glycine and the glycine
resonances in complex 2 at the same pD value. While
there is no doubt that in the solid state the complex
crystallises with the amino acid directly coordinated to
the metal [12], our pD titration experiments plus binding
and competition studies described below are consistent
with only a weak interaction between the glycines and the
metal centre in aqueous solution.
2.2. Solubility properties
Titanocene dichloride 1 has limited solubility in water,
but sonication in water results in complete dissolution
after 5–10 min to give a clear yellow solution. In
comparison, introduction of the two methyl groups to
give derivative 3, not surprisingly, reduces the aqueous
solubility significantly and sonication for 2–3 h is re-
quired to give a saturated yellow solution that frequently
contains undissolved material. In the case of 4, the
charged glycine ligands greatly improve aqueous solubil-
ity and the complex dissolves immediately in water and
no sonication is required. Replacement of the halide

G. Mokdsi, M.M. Harding / Journal of Organometallic Chemistry 565 (1998) 29–35
31
Table 1
Hydrolysis of aromatic rings in complexes 1–4 at various pD values over 24 h
Complex
X=Cl
pDa
X=gly
pDa
Time (h)b
% Precipitated
% Hydrolysis
Time (h)b
% Precipitatedc
% Hydrolysis
Cp2TiX2
2.0–2.1
6.1–6.2
\6.4
1
–
–
–
–
0
3.0–3.1
5.8–6.3
\6.4
1
0
4
60
0
24
0.25
24
–
B2d
15
24
0.25
24
–
B2d
2
X=Cl
1
X=O2CCH2NH3Cl
\90e
–
\90
\90e
–
\95f
\95f
2
(CH3Cp)2Ti
1.8–2.0
5.9–6.0
7.5–7.9
1
24
0.25
24
0.25
24
–
–
–
–
–
–
0
2.9–3.1
5.7–6.0
7.5–7.8
1
24
0.25
24
0.25
24
0
0
B2d
0
12
21
56
43
88
B2d
0
X=Cl
3
X=O2CCH2NH3Cl
4
B2d
B2d
B5g
B2d
B2d
B5g
a For pD 1.8–3.1: pD values were monitored with time and did not drop by more than 0.3 units over 24 h. For pD\5.4: pD values were
monitored with time and adjusted to the required pD range prior to recording NMR spectrum; a drop of pD unit over 24 h was typical.
b Time=0 h refers to when all sample had dissolved and pH was adjusted to required value.
c Calculated on the ratio of bound Cp signal to CH2 signal of glycine when compared to the ratio at t=1 h.
d Estimated by integration of the signals from protonated Cp or MeCp dimer/monomer; B2 when the amount of signal was too small to be
integrated accurately.
e Hydrolysis was estimated to be \90% since the amount of bound Cp signal was negligible.
f Above pD values of about 6.4 complexes 1 and 2 were hydrolysed and precipitated completely giving no signal in the NMR and so the %
precipitated was estimated at \95%.
g Estimated at B5% based on the amount of dimer and monomer peaks.
The rate of Cp hydrolysis in 1–4 was measured using
1H-NMR spectroscopy at different pD values and times
in D2O at 25°C (Table 1). In the case of 1 and 2, the rate
of Cp ring hydrolysis at three different pD values was
very similar. While hydrolysis of both 1 and 2 was
accompanied by formation of a minor amount of insol-
uble precipitate at pD ca. 2, precipitation increased at
higher pD values with \90% of the metallocene hy-
drolysed at pD 7.0. In the case of 1, the amount of
precipitate (and hence hydrolysis) could only be esti-
mated visually or by isolation and weighing the precip-
itate. However, in the case of 2, a more accurate estimate
was possible by integration of the glycine CH2 peak
versus the cyclopentadienyl protons (Table 1). Fully
dissolved complex 2 gives by integration a ratio of
10:4=Cp:CH2. As precipitation occurred the Cp signal
decreased relative to the glycine peak consistent with
formation of hydrolysis products derived from the Cp
rings.
bound MeCp multiplet (6.4 ppm at pD 7.4) and the new
methyl resonances (ca. 2 ppm) of the spectrum. As
expected, the halide ligand had little effect on the
hydrolysis experiments and similar results were obtained
for both complexes 3 and 4 at similar pD values (Table
1). At low pD values a minor amount of precipitate for-
med with time; the amount of precipitate increased when
the solution pD was raised to ca. 7, and was estimated
as described above (Table 1). NMR analysis of the
supernatant showed that the MeCp rings remained metal
bound. The yellow precipitated solid was isolated but due
to poor solubility could not be fully characterised.
The hydrolysis studies of both 3 and 4 are consistent
with loss of the halide and glycine ligands respectively in
water to give the same species, in which both MeCp rings
are bound to the metal, i.e. ‘(MeCp)2Ti2+’. In order to
confirm this interpretation, a mixed experiment was
carried out. Equimolar amounts of 3 and 4 were dis-
solved and spectra recorded at different pD values. In all
cases only one set of MeCp signals was observed consis-
tent with formation of the same (or very similar) species
in solution, presumably aquated species such as
(MeCp2)Ti(H2O)22+or (MeCp2)Ti(H2O)x(OH)(y2−y)+at
pH 7.0. A similar result was obtained on a mixed
experiment with 1 and 2.
The effect of the methyl group on the rate of ring
hydrolysis was determined by carrying out similar exper-
iments on 3 and 4. An estimate of the MeCp hydrolysis
was calculated from the relative intensities of the new
multiplets due to methylcyclopentadiene (6.5 and 6.6
ppm at pD 7.4) which appear downfield of the metal-

32
G. Mokdsi, M.M. Harding / Journal of Organometallic Chemistry 565 (1998) 29–35
2.4. Nucleotide binding studies
concomitant precipitation occurs and only signals due
to the nucleotide (i.e. Fig. 2a) are detected. In contrast
to this result, clear evidence for formation of stable
adducts between (MeCp)2TiX2 at pD ca. 7.5 was ob-
tained (Fig. 2). Fig. 2b shows the spectra obtained on
titration of 1 equiv. of 5%-dAMP into a solution of
(MeCp)2TiX2 followed by immediate adjustment of the
solution pD to 7.9 (X=Cl) and 7.7 (X=
O2CCH2NH3Cl). Fig. 2c shows the NMR spectra ob-
tained on the same samples after 24 h. Raising the
solution pD from 4 to 7.5 is accompanied by some
precipitation, but distinct signals assigned to metal-
locene complexes are clearly present at pD ca. 7.5 (Fig.
2b). The relative amount of complex is however, re-
duced to that present at lower pD values (Table 2).
After 24 h, the amount of hydrolysis increased, as
evidenced by increased precipitation, and the relative
amount of complex(es) was reduced (Fig. 2c). As was
noted for the spectra in Fig. 1, the slight differences
The major aim of this work was to prepare water
soluble, hydrolytically stable titanocene derivatives that
interact with nucleic acid constituents in the same way
as titanocene dichloride, and hence may be studied (and
administered) in preparations at pH 7.0. To this end,
nucleotide binding studies have been carried out with 3
and 4, and the stabilities, and types of complexes
formed with nucleotides compared with those formed
by 1 and 2.
Complexes 3 and 4 were used in separate titration
experiments with the nucleotides 5%-dAMP (Figs. 1 and
2) and 5%-dGMP (spectra not shown) at different pD
values, using procedures previously described [7]. In
each case, new signals were detected, and the percent-
age complexation was determined by integration of the
resonances in the region 7.5–9 ppm corresponding to
either the purine or pyrimidine protons at either pD 4
or pD 7 (Figs. 1 and 2, Table 2).
Fig. 1 compares the complexes formed between 5%-
dAMP and 1 equiv. of either 1 or 2 (Cp2TiX2; Fig. 1b)
and 5%-dAMP plus
1 equiv. of either 3 or 4
[(MeCp)2TiX2; Fig. 1c). In all four spectra, signals
assigned to uncomplexed nucleotide (Fig. 1a) are
present (indicated by dashed lines), and new signals
assigned to the formation of metallocene-nucleotide
complex(es) are detected. 31P-NMR spectra of
(MeCp)2TiCl2 (not shown) also showed the appearance
of three new signals, consistent with interaction of the
metal centre with the phosphate(O) of the nucleotide.
While the relative intensities of the signals due to the
complex(es) in the 1H-NMR spectra (Fig. 1) are slightly
different in each spectrum, when one takes into account
the slight variations in solution pD, the spectra are
remarkably similar. The effect of the methyl groups on
the Cp rings at this pD value (Fig. 1c) appears to
slightly reduce the amount of complexation compared
with Cp2TiX2 (Fig. 1b). However, the presence of the
same number of new signals at the same chemical shift
values, strongly supports the formation of the same
types of complex(es) in each case.
The effect of the X ligand on complexation is appar-
ent by comparison of the pair of spectra in either Fig.
1b or in Fig. 1c. Very similar spectra were obtained at
each pD value consistent with the conclusions of our
hydrolysis studies that indicate that in solution, both
the chloride and glycine ligands dissociate to generate
+
either ‘Cp2Ti2
’
in the case of
1 and 2, and
‘(MeCp)2Ti2+’ in the case of 3 and 4.
Fig. 2. shows the effect of increasing the solution pD
to ca. 7.4 on the relative stabilities of the complexes
formed by 3 and 4 at low pD. Spectra are not presented
for Cp2TiX2, since as reported previously for titanocene
dichloride 1, and also observed by us for 2, as soon as
the pD is raised above 5.4, complete hydrolysis and
Fig. 1. Aromatic region of 1H-NMR spectra (200 MHz, D2O) of (a)
5%-dAMP, pD 4.9; (b) 5%-dAMP+1 equiv. (MeCp)2TiX2 t=30 min,
pD 4.9 (X=Cl), 4.7 (X=gly); (c) 5%-dAMP+1 equiv. Cp2TiX2,
t=30 min, pD 5.0 (X=Cl), 4.0 (X=gly).

G. Mokdsi, M.M. Harding / Journal of Organometallic Chemistry 565 (1998) 29–35
33
Analogous studies with 5%-dUMP and 5%-dCMP were
carried out, but due to signal overlap between the
pyrimidine protons from the complex(es) and unbound
nucleotide, accurate integration and assignment of sig-
nals was not possible. For this reason, quantitative data
for the pyrimidines are not included in Table 2. The
exact structures of the species formed in solution by
(MeCp)2TiX2 and the nucleotides 5%-dAMP and 5%-
dGMP cannot be unambiguously assigned from the
NMR data. However, comparison of the chemical
shifts of the new aromatic resonances observed in the
1H-NMR spectra with those observed previously with
Cp2TiX2 [7], as well as new signals in the 31P-NMR
spectra, are consistent with phosphate(O) complexation
along with base(N) coordination and the complexes are
most probably analogous to the well-characterized Mo-
adducts (e.g. Fig. 3) [7,15]. The methyl groups do not
appear to impede complex formation or alter the bind-
ing mode compared to titanocene dichloride 1.
2.5. Conclusions
The rate of hydrolysis of the aromatic rings in
(MeCp)2TiX2 (X=Cl, O2CCH2NH3Cl) in water at pH
7 is significantly slower compared to the corresponding
unsubstituted metallocenes Cp2TiX2. While methyl sub-
stitution of the Cp rings decreases aqueous solubility,
replacement of the halide ligands with charged glycine
ligands gives the highly water soluble complex 4 in
which the Cp rings remain metal bound at physiologi-
cal pH, and which forms complexes with nucleotides
which are stable for 24 h at pD 7.4. Substitution of
alkyl groups on titanocene dichloride has been reported
to generally decrease the antitumour activity of the
drug [16]. However, the reduced aqueous solubility of
the complexes as well as administration of the com-
pounds in DMSO/saline at low pH needs to be taken
into account in the interpretation of biological data.
While the glycine ligands present in both 2 and 4
improve aquoues solubility, these ligands may affect the
rate of cellular uptake of the metallocenes. Screening of
Fig. 2. Aromatic region of 1H-NMR spectra (200 MHz, D2O) of (a)
5%-dAMP, pD 7.7; (b) 5%-dAMP+1 equiv. (MeCp)2TiX2 t=30 min,
pD 7.9 (X=Cl), 7.7 (X=gly); (c) 5%-dAMP+1 equiv. (MeCp)2TiX2,
t=24 h, pD 7.8 (X=Cl), 7.5 (X=gly).
between the relative intensities of the signals in the
spectra are within the experimental limits of fluctuation
of pD with time.
Similar trends were observed with 5%-dGMP (spectra
not shown), with the results summarised in Table 2.
Table 2
Percentage complexation between Cp2TiX2 and (MeCp)2TiX2 and nucleotidesa
Complex
+1 eq. Nucleotide
X=Cl
X=gly
t (h)
pD
4.9
7.9
7.8
4.7
10.1
7.5
% Complex
t (h)
pD
% Complex
(MeCp)2TiX2 X=Cl 3
X=O2CCH2NH3Cl 4
5%-dAMP
5%-dGMP
0.5
0.5
24
0.5
0.5
24
70
62
40
60
14
14
40
45
0.5
0.5
24
0.5
0.5
24
4.7
7.7
7.5
3.4
7.8
7.5
4.0
4.6
60
50
30
60
30
20
50
55
Cp2TiX2 X=Cl 1
X=O2CCH2NH3Cl 2
5%-dAMP
5%-dGMP
0.5
0.5
5.0
5.0
0.5
0.5
a Determined by integration of purine aromatic resonances in 1H-NMR spectra.

34
G. Mokdsi, M.M. Harding / Journal of Organometallic Chemistry 565 (1998) 29–35
mixture was stirred for 5 h at r.t., and filtered through
a U-tube containing glass wool. A sample of the filtered
reaction mixture containing sodium methylcyclopenta-
diene (14 ml, 36 mmol, 2.6 M) was added dropwise to
titanium(IV) tetrachloride (1 ml, 9 mmol) in THF (40
ml) at −20°C. The reaction mixture was stirred
overnight at r.t., the solvents were removed and the
brown–red residue was dissolved in chloroform (70 ml)
and stirred with flash silica for 40 min. The reaction
mixture was filtered and the residual flash silica was
further extracted with chloroform (2×100 ml). The
solvent was removed from the dark, blood-red filtrate
to give a brown–red precipitate which was dried under
vacuum. The crude product was recrystallised several
times from chloroform to give bis(methylcyclopentadi-
enyl) titanocene dichloride 3 (350 mg, 14%) as a red
diamond shaped crystalline solid, m.p. dec. \200°C
(lit. [13] dec. 217–218°C). Anal. Calcd for
(MeCp)2TiCl2: C, 52.0; H, 5.1; Cl, 25.6. Found: C, 52.3;
Fig. 3. Tentative assignment of possible complexes formed between
(Cp2Me)2TiX2 and 5%-dAMP at pH 7.
complex 4, which may be administered in aqueous
solution at pH 7.4, and is capable of formation of
stable adducts with nucleotides may provide further
insight into the mechanism of antitumour action of
titanocene based metallocenes.
1
H, 5.0; Cl, 25.5. H-NMR (D2O, pD 1.5) l 6.3–6.4 (m,
3. Experimental
8H, Cp), 2.3 (s, 6H, Me). EI MS m/z 276
[(MeCp)2TiCl2+, 29%]; 241 [(MeCp)2TiCl+, 31]; 204
[(MeCp)2Ti+, 69]; 197 [(MeCp)TiCl2+, 97]; 161
[(MeCp)TiCl+, 100]; 126 [(MeCp)Ti+, 37].
3.1. General
Nucleotides were purchased from Sigma Chemical
Company and were used as provided. Titanocene
dichloride 1 was obtained from the Aldrich Chemical
Company. The bisglycine analogue 2 was prepared
from titanocene dichloride 1 according to the literature
procedure [12]. Melting points were determined on a
Reichert heating stage and are uncorrected. Mass spec-
tra were recorded on an A.E.I. MS-902 spectrometer at
70 eV. Values of m/z are quoted with intensities ex-
pressed as percentages of the base peak in parentheses.
1H-NMR spectra were recorded on a Bruker AC200
NMR spectrometer and were referenced to TSP (0.00
ppm). 31P-NMR spectra were recorded on a Bruker
AMX400 spectrometer (161.98 MHz) and were refer-
enced to external neat trimethyl phosphite (140.85 ppm)
3.3. Preparation of bis(methylcyclopentadienyl)
titanocene bis(glycine) 4
Bis(methylcyclopentadiene) titanocene dichloride 3
(280 mg, 1.0 mmol) and glycine (150 mg, 2.0 mmol)
were stirred in methanol (20 ml) at r.t. under a nitrogen
atmosphere for 4 h. The reaction mixture was filtered,
and the filtrate was evaporated to dryness. The
semicrystalline product solidified after 1 day at −17°C.
The solid was washed with chloroform and dried under
vacuum to give bis(methylcyclopentadienyl) titanocene
bis(glycine) 4 (340 mg, 75%) as an orange crystalline
solid, m.p. dec. \200°C. Anal. Calcd for
(MeCp)2Ti(OOCCH2NH3Cl)2 ·1.5H2O: C, 42.3; H, 6.0;
1
N, 6.2. Found: C, 42.1; H, 5.8; N, 6.4. H-NMR (D2O,
3.2. Preparation of bis(methylcyclopentadiene)
titanocene dichloride 3
pD 2.5) l 6.4 (m, 8H, MeCp), 3.68 (s, 4H, CH2), 2.01
(s, 6H, Me).
Methylcyclopentadiene dimer (200 ml) was cracked
at \200°C over a 20 cm column of glass beads to give
methylcyclopentadiene (b.p. 72°C) which was collected
at −78°C, recracked and redistilled several times
through a 40 cm column of glass beads, and stored at
−78°C until used. The initial low boiling fractions
(B70°C) were discarded. Sodium sand was generated
by refluxing sodium (4.7 g, 0.20 mol) in toluene (100
ml). The toluene was removed, and the sodium sand
was washed with THF (3×25 ml) and THF (50 ml)
was added. Methylcyclopentadiene (13 ml, 10.4 g, 0.13
mol) was added dropwise over 5 min to the suspension
of sodium sand at −10 to −15°C. The reaction
3.4. Hydrolysis experiments
The general procedure involved dissolving 5–15 mmol
of the complex in 500 ml D2O with TSP added for
reference. Dissolution of 3 required sonication for 200
min to give a clear yellow solution. Complex 4 was
dissolved by shaking for 2 min. Sonication was carried
out with an Elma Transsonic Digital Bath. The solution
pD was adjusted with DCl and NaOD. pD values were
measured using a Beckman F11 meter and a Mettler
NMR tube pH probe and are related to the pH meter
reading by the formula pD=pH(meter reading)+0.4
[17]. Measured pD values are 90.3 due to fluctuations

G. Mokdsi, M.M. Harding / Journal of Organometallic Chemistry 565 (1998) 29–35
35
1
[3] L.Y. Kuo, A.H. Liu, T.J. Marks, Met. Ions Biol. Syst. 33 (1996)
53.
in sample pH which occurred over 30 min. H-NMR
spectra were recorded at time intervals with any devel-
oping precipitate ignored. The rate of Cp hydrolysis
was estimated by integration of one of the two multi-
plets arising from free cyclopentadiene C5H5D (6.5 and
6.6 ppm at pD=6.2) versus the metal-bound C5H5
signals.
[4] L.Y. Kuo, M.G. Kanatzidis, M. Sabat, L. Tipton, T.J. Marks,
Am. Chem. Soc. 113 (1991) 9027.
[5] M.M. Harding, G.J. Harden, L.D. Field, FEBS Letts. 322 (1993)
291.
[6] M.L. McLaughlin, J.M. Cronan Jr., T.R. Schaller, R.D. Sneller,
J. Am. Chem. Soc. 112 (1990) 8949.
[7] J.H. Murray, M.M. Harding, J. Med. Chem. 37 (1994) 1936.
[8] J.H. Toney, T.J. Marks, J. Am. Chem. Soc. 107 (1985) 947.
[9] See for example, (a) P. Ko¨pf-Maier, T. Klapo¨tke, H. Ko¨pf, F.
Preiss, T. Marx, Anticancer Res. 6 (1986) 33. (b) P. Ko¨pf-Maier,
T. Klapo¨tke, Arzneim-Forsch. 37 (1987) 532. (c) P. Ko¨pf-Maier,
Eur. J. Clin. Pharmacol. 47 (1994) 1.
Acknowledgements
The award of an Australian Postgraduate Award
(G.M.) is gratefully acknowledged.
[10] P. Ko¨pf-Maier, B. Hesse, R. Voigtlander, H. Ko¨pf, J. Cancer
Res Clin Oncol. 97 (1980) 31.
[11] J.L. Robbins, N. Edelstein, B. Spencer, J.C. Smart, J. Am.
Chem. Soc. 104 (1982) 1882.
References
[12] (a) T.M. Klapo¨tke, H. Ko¨pf, I.C. Tornieporth-Oetting, P.S.
White, Organometallics 13 (1994) 3628. (b) T.M. Klapo¨tke, H.
Ko¨pf, I.C. Tornieporth-Oetting, P.S. White, Angew. Chem. Int.
Ed. Engl. 33 (1994) 1518.
[13] L.T. Reynolds, G.J. Wilkinson, Inorg. Nucl. Chem. 9 (1959) 86.
[14] P. Ko¨pf-Maier, S. Grabowski, J. Liegener, H. Ko¨pf, Inorg.
Chim. Acta 108 (1985) 99.
[1] For reviews see (a) P. Ko¨pf-Maier, H. Ko¨pf, Drugs of the
Future 11 (1986) 297. (b) P. Ko¨pf-Maier, H. Ko¨pf, Struct. Bond.
70 (1988) 105. (c) P. Ko¨pf-Maier, H. Ko¨pf, in: S.P. Fricker (Ed.),
Metal Compounds in Cancer Chemotherapy, Chapman & Hall,
London, 1994, p. 109.
[15] L.Y. Kuo, M.G. Kanatzidis, M. Sabat, L. Tipton, T.J. Marks, J.
Am. Chem. Soc. 113 (1991) 9027.
[16] P. Ko¨pf-Maier, W. Kahl, N. Klouras, G. Hermann, H. Ko¨pf,
Eur. J. Med. Chem. 16 (1981) 275.
[2] (a) W.E. Berdel, H.J. Schmoll, M.E. Scheulen, A. Korfel, M.F.
Knoche, A. Harstrci, F. Bach, J. Baumgart, G.J. Sass, Cancer
Res. Clin. Oncol. 120 (Suppl.) (1994) R172. (b) V.J. Moebus, R.
Stein, D.G. Kieback, I.B. Runnebaum, G. Sass, R. Kreienberg,
Anticancer Res 17 (1997) 815.
[17] P.K. Glasoe, F.A. Long, J. Phys. Chem. 64 (1960) 188.
.
.