Synthesis, characterization, cytotoxicity, and hydrolytic behavior of C2- and C1-symmetrical TiIV complexes of tetradentate diamine bis(phenolato) ligands: A new class of antitumor agents

Article abstract of DOI:10.1002/chem.200801310

We recently introduced a new class of bis(isopropoxo)-Ti^IV complexes with diamine bis(phenolato) ligands that possess antitumor activity against colon HT-29 and ovarian OVCAR-1 cells that is higher than that of the known Ti^IV compoun

Full text of DOI:10.1002/chem.200801310

FULL PAPER
DOI: 10.1002/chem.200801310
Synthesis, Characterization, Cytotoxicity, and Hydrolytic Behavior of C2- and
C1-Symmetrical TiIV Complexes of Tetradentate Diamine Bis(Phenolato)
Ligands: A New Class of Antitumor Agents
Dani Peri, Sigalit Meker, Michal Shavit, and Edit Y. Tshuva*[a]
Abstract: We recently introduced a
new class of bis(isopropoxo)–TiIV com-
plexes with diamine bis(phenolato) li-
gands that possess antitumor activity
against colon HT-29 and ovarian
OVCAR-1 cells that is higher than that
of the known TiIV compounds titano-
cene dichloride and budotitane as well
as that of cisplatin. Herein, we elabo-
rate on this family of compounds; we
discuss the effect of structural parame-
ters on the cytotoxic activity and hy-
drolytic behavior of these complexes,
seeking a relationship between the two.
Whereas complexes with small steric
groups around the metal center possess
high activity and lead mostly to forma-
tion of O-bridged polynuclear com-
plexes with bound bis(phenolato)
ligand upon water addition, bulky com-
plexes hydrolyze to release all free li-
gands and are inactive. Slightly increas-
ing the size of the N-donor substituents
probably weakens the ligand binding in
solution, and, thus, rapid hydrolysis is
observed, leading to a lack of cytotox-
icity, supporting the requirement for
ligand inertness. Replacing the two iso-
propoxo ligands with a single catecho-
lato unit gives a complex with a differ-
ent geometry that exhibits slower hy-
drolysis and reduced cytotoxicity, sug-
gesting some participation of labile
ligand hydrolysis in the cytotoxicity
mechanism.
A
crystallographically
characterized O-bridged polynuclear
species obtained from a biologically
active bis(isopropoxo) complex upon
water addition is inactive, which rules
out its participation as the active spe-
cies, yet suggests some role of the par-
ticular steric and electronic require-
ments allowing its formation in the ac-
tivity mechanism. Additional measure-
ments support rapid formation of the
active species in the presence of cells
prior to O-bridged TiIV cluster forma-
tion.
Keywords: antitumor agents · cyto-
toxicity
· hydrolysis · phenolato
ligands · titanium
Introduction
various reactivities and mechanisms that may lead to im-
proved anticancer therapeutics.[1–13] The TiIV complexes tita-
nocene dichloride ([TiCl2Cp2], Cp=cyclopentadienide) and
In the last two decades, much focus has been given to non-
platinum transition-metal complexes that may exhibit cyto-
toxic properties in attempts to discover novel reagents of
budotitane ([TiACHTUNTRGENNUG(bzac)2ACHUTNGTREN(NUGN OEt)2], bzac=benzoylacetonate,
OEt=ethoxide) are two parent compounds (shown here)
that possess promising antitumor activity against cisplatin-
sensitive and -resistant cells; however, their hydrolytic insta-
bility inhibited their applicable use.[4–6,14–20] In the presence
of water, the labile groups (Cl, OR; OR=alkoxide) are hy-
drolyzed first within seconds, followed by the hydrolysis of
the more inert ligands (Cp, diketonato) within hours,[21]
leading to unidentified aggregates that hamper further
[a] D. Peri, S. Meker, M. Shavit, Dr. E. Y. Tshuva
Institute of Chemistry
The Hebrew University of Jerusalem, Jerusalem 91904 (Israel)
Fax : (+972)2-658-4282
E-mail: tshuva@chem.ch.huji.ac.il
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.200801310. It contains color
graphics for Figure 4–7, Figure 10–12, Figure 16, Figure 17; cytotoxici-
ty measurements on OVCAR-1 cells, overlay of UV/Vis spectra upon
water addition of [Ti(Ln)
after D2O addition of [Ti(Ln)
RT, NMR spectra for 14 h following D2O addition at 378C for
[Ti(L2)(OiPr)2], NMR integration plots versus time for [Ti(Ln)-
(OiPr)2] (n=2–4) at RT and/or at 378C, and spectroscopic data on
[Ti3(L1)
3A(m2-O)3].
ACHTUNGTRENNUNG
A
ACHTUNGTRENNUNG
ACHTUNGTRENNUNG
ACHTUNGTRENNUNG
CHTUNGTRENNUNG
Chem. Eur. J. 2009, 15, 2403 – 2415
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
2403

mechanistic investigations and ultimately lead to their fail-
ure in clinical trials. The nature of the active species, there-
fore, remains unknown, as does the identity of the biological
target, the nature of the participation of the ligands, and the
role of their particular lability.[20,22–25] Thus, the development
of new TiIV-based cytotoxic compounds of defined hydrolytic
behavior that may shed some light on this enigmatic process
remains a significant challenge
Results and Discussion
Ligand selection and complex preparation: In the current
study, we were interested in evaluating the effect of structur-
al parameters on the cytotoxic activity of this family of dia-
mine–bis(phenolato) TiIV complexes and in gaining informa-
tion on their hydrolytic behavior. Steric effects were exam-
of immense value.
Another interesting observa-
tion is the interaction of titano-
cene dichloride with the serum
protein transferrin.[26–31] This
protein leads to complete
ligand stripping from the TiIV
core, and apparently allows
somewhat selective cell pene-
tration through the transferring
receptors located more abun-
dantly on tumor cell surfaces.
Nevertheless, early loss of the
most inert ligands abolishes
their potential influence on the
interaction with the biological
target, and, thus, questions the
contribution and influence of
the particular ligand on the cy-
totoxic activity. Recent studies,
however, suggested that poten-
tial binding of TiIV to albumin
may also be achieved in the
form of a ligand-bound complex,[32] which restores the inter-
est in ligand design and structure–activity relationship stud-
ies by ligand modification, especially since labile complexes
ined by incorporating different substituents on the aromatic
rings and N-donors. Particularly, the location of large steric
groups may be of importance to elucidate whether their
effect results from proximity to the metal and thus interferes
with its interaction with various targets throughout hydroly-
sis or biological activity or from disruption to planarity
which should preclude intercalation to DNA, a potential
biological target, as has been suggested for budotitane.[17]
such as [Ti
ACHTUNGTRENNUNG(OiPr)4] and [TiCl4ACHUTNGTRENN(UGN thf)2] are biologically inac-
tive.[23]
We have recently introduced a new family of cytotoxic
agents based on readily accessible TiIV complexes of diamine
bis(phenolato) ligands.[33] These complexes demonstrate
some resemblance to the active budotitane in terms of cova-
lent donor types, coordination number, and general symme-
try; however, unlike budotitane, they are obtained quantita-
tively in a single step as single (racemic) isomers[17,34] and in-
clude a single tetradentate dianionic ligand instead of two
monoanionic ones. Some members of this C2-symmetrical
The complexes [Ti(Ln)
ACTHNUTRGNEN(UG OiPr)2] (n=1–5) were prepared in
quantitative yields according to a highly convenient, pub-
lished procedure from the ligand precursors H2Ln (n=1–5)
and [TiACTHNURTGENNG(U OiPr)4] at room temperature in diethyl ether or
THF.[33,35–37] The ligands were all synthesized by a single-step
Mannich condensation from the substituted phenol, formal-
dehyde, and the N,N’-disubstituted diamine.[38] As we recent-
family of complexes ([Ti(Ln)
ACTHNUTRGNEUNG(OiPr)2] n=1 or 2) exhibit
higher activity against colon HT-29 and ovarian OVCAR-1
cells than those measured for titanocene dichloride, budoti-
tane, and cisplatin, which is independent of transferrin.[32]
The reactivity is, however, strongly dependent on the partic-
ly communicated,[33] the X-ray structure of [Ti(L1)
ACHTUNGTRENNUNG(OiPr)2]
(Figure 1, top; Table 1) indicates a C2-symmetrical complex
with two cis isopropoxo groups and with the phenolato
groups in a trans configuration, similar to other known com-
ular ligand employed, since bulkier ligands ([Ti(L3)
N
pounds of this class.[36,37,39,40] Complex [Ti(L4)
ACTHNGUTERN(NUG OiPr)2] was
aliphatic ligands, and ligands of branched donor connectivity
that produce complexes of different symmetry all lead to
substantially reduced activity.[33] Herein we elaborate on this
family of complexes, discussing the effect of structural pa-
rameters on cytotoxicity and hydrolytic behavior of the com-
plexes, while pointing to a connection between the two.
also crystallized from diethyl ether by slow evaporation at
room temperature and the X-ray structure (Figure 2;
Table 2) reveals a similar general geometry of a C2-symmet-
rical complex with two cis isopropoxo groups and a trans
configuration of the phenolato oxygen donors. Complex
[Ti(L5)
(OiPr)2] was crystallized from diethyl ether at À358 C
2404
www.chemeurj.org
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 2403 – 2415

TiIV Complexes
FULL PAPER
Table 2. Selected bond lengths [ꢁ] and angles [8] for [Ti(L4)
ACTHNUTRGNEU(GN OiPr)2].
À
À
O(1) Ti
1.900(2)
1.905(2)
1.820(2)
1.816(2)
N(1) Ti
2.312(3)
2.360(3)
À
À
O(2) Ti
N(2) Ti
À
O(3) Ti
À
O(4) Ti
O(4)-Ti-O(3)
O(4)-Ti-O(1)
O(3)-Ti-O(1)
O(4)-Ti-O(2)
O(3)-Ti-O(2)
O(1)-Ti-O(2)
N(1)-Ti-N(2)
104.01(12)
91.25(11)
98.31(11)
96.53(11)
93.29(11)
164.06(10)
76.26(12)
O(4)-Ti-N(1)
O(3)-Ti-N(1)
O(1)-Ti-N(1)
O(2)-Ti-N(1)
O(4)-Ti-N(2)
O(3)-Ti-N(2)
O(1)-Ti-N(2)
O(2)-Ti-N(2)
164.88(12)
89.92(12)
80.79(10)
88.36(10)
90.40(11)
164.94(11)
85.52(10)
80.55(10)
Figure 3. ORTEP drawing of [Ti(L5)
soids; H atoms omitted for clarity.
ACHTUNGTNER(NUNG OiPr)2] with 50% probability ellip-
Figure 1. ORTEP drawing of [Ti(L1)(OiPr)2] (top) and [Ti(L1)
ACTHNUTRGENNUG ACHTUNGTREN(NUGN O2Ph)]
(bottom) with 50% probability ellipsoids; H atoms omitted for clarity.
Table 3. Selected bond lengths [ꢁ] and angles [8] for [Ti(L5)
ACTHNUTRGNEU(GN OiPr)2].
À
À
O(1) Ti
1.904(2)
1.898(2)
1.835(2)
1.814(2)
N(1) Ti
2.369(2)
2.376(2)
Table 1. Selected bond lengths [ꢁ] and angles [8] for [Ti(L1)
ACHTUNGTRENNUNG
À
À
O(2) Ti
N(2) Ti
À
O(3) Ti
À
À
O(1) Ti
1.9001(11)
1.8208(13)
N(1) Ti
À
O(4) Ti
À
O(2) Ti
O(4)-Ti-O(3)
O(4)-Ti-O(1)
O(3)-Ti-O(1)
O(4)-Ti-O(2)
O(3)-Ti-O(2)
O(1)-Ti-O(2)
N(1)-Ti-N(2)
105.81(10)
92.01(9)
95.22(9)
95.24(9)
92.45(9)
167.61(9)
75.49(8)
O(4)-Ti-N(1)
O(3)-Ti-N(1)
O(1)-Ti-N(1)
O(2)-Ti-N(1)
O(4)-Ti-N(2)
O(3)-Ti-N(2)
O(1)-Ti-N(2)
O(2)-Ti-N(2)
163.49(9)
90.28(9)
82.68(8)
87.58(9)
88.91(9)
164.24(9)
89.71(8)
80.41(8)
O(2)-Ti-O(2)
O(2)-Ti-O(1)
O(2)-Ti-O(1)
O(1)-Ti-O(1)
N(1)-Ti-N(1)
106.10(10)
96.85(6)
90.89(5)
167.13(7)
75.84(9)
O(1)-Ti-N(1)
O(1)-Ti-N(1)
O(2)-Ti-N(1)
O(2)-Ti-N(1)
88.61(5)
81.22(5)
163.37(7)
89.44(6)
5
À
ticular, the Ti N distances in [Ti(L )
G
are only slightly longer than those observed for [Ti(L1)-
AHCTUNGTRENNUNG
extra methylene units on the complex structure and geome-
try is observed (Table 1, Table 3).
To evaluate the role of the lability of the OiPr
groups,[20,22–24] we also prepared the complex [Ti(L1)
ACHTUNGTRENNUNG(O2Ph)]
(Ph=phenyl), which has one bidentate catecholato ligand
Figure 2. ORTEP drawing of [Ti(L4)
soids; H atoms omitted for clarity.
ACHTUNGTNER(NUNG OiPr)2] with 50% probability ellip-
instead of the two monodentate isopropoxo groups. This
complex was synthesized by reacting [Ti(L1)
ACTHUNTRGNE(UGN OiPr)2] with
and its X-ray structure (Figure 3; Table 3) again shows a C2-
symmetrical complex with two cis isopropoxo groups and
two trans phenolato donors, which is of very high structural
one equivalent of ortho-catechol in THF at room tempera-
ture (Scheme 1). The H NMR spectrum of the product re-
vealed low symmetry, with four aromatic singlets of the bis-
(phenolato) ligand, four doublets of two AX systems of the
ArCH2 (Ar=aryl) groups, four doublets-of-triplets or mul-
1
similarity to its dimethyl analogue [Ti(L1)
ACHTUNGTRENNUNG
Chem. Eur. J. 2009, 15, 2403 – 2415
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
2405

E. Y. Tshuva et al.
Scheme 1. Replacing two isopropoxo groups with a single catecholato
unit.
tiplets of the NCH2CH2N bridge, four different Ar-CH3 sin-
glets, two different NMe signals, and two doublets-of-dou-
blets and two doubles-of-triplets of the catecholato unit.
Single crystals of [Ti(L1)
ACTHNUTRGNENUG(O2Ph)] suitable for X-ray crystal-
lography were grown from diethyl ether at room tempera-
ture, and the ORTEP drawing of the structure is presented
in Figure 1 (bottom). A list of selected bond lengths and
angles is given in Table 4.
Figure 4. Dependence of HT-29 cell viability after 3 days of incubation
on the concentration of [Ti(Ln)
ACTHUNRGTENUNG(OiPr)2] (n=1–3) and [TiCl2Cp2] adminis-
tered presented in logarithmic scale with or without added apo-transfer-
rin (10 mgmLÀ1); see IC50 values in reference.[33] Thick solid curves:
3
2
~
^
: [Ti(L )
ACHTUGTNERNNUG(OiPr)2]+apo-transferrin; : [Ti(L )ACHUTNGTREN(NUNG OiPr)2]+apo-transferrin;
&
*
:
:
[Ti(L1)
Other curves: : [Ti(L )
(OiPr)2]+apo-transferrin;
[TiCl2Cp2]+apo-transferrin;
(OiPr)2]; : [Ti(L )
3
2
1
Table 4. Selected bond lengths [ꢁ] and angles [8] for [Ti(L1)
ACHTUNGTRENNUNG
~
&
^
(OiPr)2]; : [Ti(L )
G
N
*
: [TiCl2Cp2].
À
À
O(1) Ti
1.835(3)
1.849(3)
1.961(3)
1.909(3)
N(1) Ti
À
À
O(2) Ti
N(2) Ti
2.300(3)
À
O(3) Ti
large steric groups. It is also clear that the cell penetration
mechanism for these complexes is independent of transfer-
rin, unlike that of titanocene dichloride and similar to that
of budotitane.[32] The dependence of the reactivity on the in-
cubation time and the total loss of activity observed after
exposing the complexes to the biological medium for two
days prior to cell addition (Figure 5) suggest that the formed
active species penetrates the cells quite rapidly through in a
non-transferrin-dependent fashion, prior to complex dissoci-
ation and activity loss, and once in the cell the reactivity is
retained.[33]
À
O(4) Ti
O(4)-Ti-O(3)
O(4)-Ti-O(1)
O(3)-Ti-O(1)
O(4)-Ti-O(2)
O(3)-Ti-O(2)
O(1)-Ti-O(2)
N(1)-Ti-N(2)
81.05(12)
107.61(12)
94.45(12)
97.86(13)
170.58(12)
94.79(12)
78.04(12)
O(4)-Ti-N(1)
O(3)-Ti-N(1)
O(1)-Ti-N(1)
O(2)-Ti-N(1)
O(4)-Ti-N(2)
O(3)-Ti-N(2)
O(1)-Ti-N(2)
O(2)-Ti-N(2)
161.02(12)
82.38(12)
82.81(12)
96.98(13)
92.19(12)
88.03(12)
160.20(13)
82.65(11)
Interestingly, the structure of [Ti(L1)
(O2Ph)] features a
To evaluate the role of the particular location of the large
tBu (tBu=tert-butyl) groups on abolishing activity, we stud-
compound of a different symmetry, namely C1-symmetrical
complex with the two phenolato groups of the tetradentate
ligand in a cis configuration (Figure 1, bottom),[39] in con-
trast to the geometry of the starting complex (Figure 1, top).
This result suggests an associative mechanism for ligand re-
placement, according to which a seven-coordinate inter-
mediate is obtained upon the first nucleophilic attack on the
metal center that causes ligand rearrangement, and a rela-
tively rapid second attack to give the final product.
ied the complex [Ti(L4)
ACTHNUTRGNEN(UG OiPr)2], which includes a single
Cytotoxicity measurements and steric effects in [Ti(Ln)-
ACHTUNGTRENNUNG(OiPr)2] (n=1–5): Cytotoxicity was measured on colon HT-
29 and ovarian OVCAR-1 cells according to the MTT
(methylthiazolyldiphenyltetrazolium bromide) assay.[33] The
results obtained for [Ti(Ln)
ACTHNUTRGEN(UNG OiPr)2] (n=1–3) in comparison
to [TiCl2Cp2] after three days of incubation of the complexes
with the cells with and without added apo-transferrin to the
medium are summarized in Figure 4.[33] Substantial activity
was observed for the complexes [Ti(Ln)
ACTHNUTRGNE(UNG OiPr)2] (n=1 or 2),
Figure 5. Dependence of HT-29 cell viability on incubation time a) with
or b) without 2 days of exposure to the biological medium prior to cell
while no activity was observed for the bulky complex
[Ti(L3)
N
administration of [Ti(Ln)
ACTHNGUTRENU(NG OiPr)2] (n=1 and 2 ) at 0.1 mm.
&
^
2406
www.chemeurj.org
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 2403 – 2415

TiIV Complexes
FULL PAPER
bulky substituent that is rather distant from the metal site.
Interestingly, some activity was observed towards HT-29 and
ment of the OiPr groups were also made by a reaction of
the complexes with specific amounts of water and crystalli-
zation of the resulting product.
OVCAR-1 cells, which is between that of [Ti(L1)
ACHTUNGTRENUNG
and that of [Ti(L3)
ACHTUNGTRENNUNG
UV/Vis measurements: In a general procedure, a TiIV com-
plex was dissolved in THF, and water was added to give a
1:9 water/THF solution. The absorbance was monitored
overtime for several hours. Interestingly, a slight shift in the
LMCT band to longer wavelengths was detected within sev-
eral hours, for the biologically active complexes [Ti(Ln)-
AHCTUNGTRENNUNG
Figure 6. Dependence of HT-29 cell viability after 3 days of incubation
on the concentration of [Ti(Ln)
sented in logarithmic scale.
ACHTUNGTNER(NUNG OiPr)2] (n=1 and 3–5) administered pre-
:
[Ti(L1)
: [Ti(L )ACHTUNGTRENNUNG(OiPr)2]; : [Ti(L )ACHTGNUTREN(NUNG OiPr)2].
G
: ACHTUNTGNUER(GN OiPr)2];
[Ti(L3)
~
&
4
5
^
*
tent with the observation regarding the negative effect of
steric groups on activity; it points to an influence of periph-
eral groups as well and does not rule out a possible role of
DNA intercalation in the cytotoxic activity of these com-
plexes, especially considering the similar activity of [Ti(L1)-
Figure 7. UV/Vis absorption over time for [Ti(L1)
of water to a solution of the complex in THF giving a 10% water
solution.
ACHTUNGTERN(NUGN OiPr)2] upon addition
AHCTUNGTRENNUNG
A
ACHTUNGTRENNUNG(OiPr)2] (Figure 4).
includes a bound bis(phenolato) ligand under these condi-
tions. No such shift was detected for the biologically inactive
AHCTUNGTRENNUNG
ethyl substituents on the bound nitrogen atoms instead of
methyl groups is completely inactive against both cell types
analyzed (Figure 6), despite its highly similar X-ray structure
ACTHNGUTRENNUG CAHTGNUTREN(NUNG OiPr)2] dem-
[Ti(Ln)(OiPr)2] (n=3 and 5), for which [Ti(L5)
onstrated a significantly more rapid hydrolysis process with
the LMCT band fully decaying within the first hour of water
addition.
to that of [Ti(L1)
ACHTUNGTRENNUNG(OiPr)2] (Figures 1 and 3). It appears that
even a rather small increase of the substituent bulk, when
employed directly on the donor to the metal, has a signifi-
cant negative influence on activity. This suggests that either
an approachable metal site or strong ligand binding are re-
As we suspected that exchange of the isopropoxo groups
with water molecules giving O-bridged species should take
place rapidly for all of these bis(isopropoxo) com-
plexes,[1,14,21] we used a stopped flow instrument to identify
possible quick changes within the first few seconds of water
À
quired for activity, as the Ti N bonds have probably been
weakened by the larger substituents, despite the rather simi-
lar bond distances observed in the solid-state structures,
which may also be affected by crystal packing parameters.
addition for [Ti(Ln)
observed.
ACHTUNGTNER(NUNG OiPr)2] (n=1–3); no such changes were
NMR measurements: To verify that no rapid hydrolysis reac-
Hydrolysis studies and steric effects in [Ti(Ln)
(OiPr)2] (n=
tion occurs within the first few minutes following water ad-
1–5): We investigated the hydrolytic behavior of the com-
plexes by three methods. Changes in the UV/Vis absorption
upon water addition, particularly of the ligand-to-metal
charge transfer (LMCT) band of the bis(phenolato) ligand
at l=320–350 nm allowed the evaluation of general hydro-
lytic behavior and the possible identification of new com-
pounds formed over time. NMR studies were valuable to
determine the particular nature of the new compounds, if
formed. In addition, attempts to intentionally synthesize par-
tial hydrolysis products with O-bridging ligands in replace-
dition for [Ti(Ln)
ACTHNGUTERN(UNG OiPr)2] (n=1–4) and to gain information
on the nature of the new species that may form within hours
in some cases, as suggested by the UV-vis measurements, we
added D2O to a solution of the complexes in [D8]THF to
1
give a 1:9 D2O/[D8]THF mixture and took a H NMR mea-
surement every 4–10 minutes for several hours. To our sur-
prise, the first spectrum taken following D2O addition did
not reveal a rapid reaction, and no indication of substantial
isopropanol formation was observed for [Ti(Ln)
ACHTUNGTRENNUNG
Chem. Eur. J. 2009, 15, 2403 – 2415
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
2407

E. Y. Tshuva et al.
began to appear (Figure 8). For [Ti(L5)
ACTHNUTRGNEU(NG OiPr)2], the isopro-
panol signals appeared much more rapidly, in the very first
measurement following D2O addition (Figure 9), consistent
with the results obtained by UV/Vis spectroscopy. Interest-
ingly, for the biologically active complexes [Ti(Ln)
ACTHNUTRGNENU(G OiPr)2]
(n=1, 2, and 4), the signals representing the free ligands
that appeared within the measurement period were quite
small and accompanied by many additional new signals of
some new species (Figure 8), which appeared more rapidly
for [Ti(Ln)
ACHTUNGTRENNUNG
for [Ti(L2)
ACHTUNGTRENNUNG
trast, only signals representing isopropanol and the free
ligand H2L3 were observed for the biologically inactive
bulky complex [Ti(L3)(OiPr)2]. Complex [Ti(L5)
ACTHNUGTRENNUGN ACHTUNGTRENNNUG
[Ti(Ln)
ACHTUNGTRENNUNG(OiPr)2] (n=1–4), also leads mainly to free ligand
H2L5 signals although some tiny additional signals may be
À
detected (Figure 9). Plotting the integration of the Ti OCH-
(CH3)2 signal, the (CH3)2CHOH signal, one signal of a
bound bis(phenolato) ligand, and one signal of a free li-
gand[41b] versus time (Figures 10–12) revealed that, for all
complexes, all of the originally bound isopropoxo ligands
are accounted for in the isopropanol formed, and in addi-
tion, the decay of integration of the bound bis(phenolato)
ligand signal occurred simultaneously with isopropanol for-
mation. Looking closely at the results obtained for the bio-
Figure 9. 1H NMR spectra of [Ti(L5)
ACTHNUGTRENNUG(OiPr)2] at RT a) in [D8]THF, b) im-
mediately following the addition of D2O, and c) 1 h following the addi-
tion of D2O.
logically inactive complexes [Ti(L3)
(OiPr)2] and [Ti(L5)-
ACHTUNGTRENNUNG
AHCTUNGTRENNUNG
bis(phenolato) ligand increases over a similar timescale to
that of isopropanol formation despite the expected differ-
ence in their lability (Figures 11 and 12);[41b] however, while
for [Ti(L3)
ACHTUNGTRENNUNG
(OiPr)2], only ꢀ50% of the bound bis(phenolato)
ligand can be accounted for (Figure 12). Although we could
not clearly detect a gradual increase in integration for the
new signals observed for [Ti(L5)
ACTHNUTRGENUN(G OiPr)2], it is conceivable
that some new species is obtained with lower symmetry,
which is represented by the multiple, small, difficult-to-ana-
lyze signals. One distinctive feature in the hydrolytic behav-
ior of [Ti(L3)
ACTHNUGTRNE(UNG OiPr)2] (Figure 11) is slower initial hydrolysis
that becomes faster with time and is probably due to the re-
lease of tension following partial ligand dissociation. As all
new signals observed for the biologically active complexes
[Ti(Ln)
ACTHNUGTRNE(UNG OiPr)2] (n=1 and 2) have quite small integration
values, increasing the temperature of the NMR experiment
to 378C was valuable to accelerate the reaction and observe
its completion upon total decay of signals relating to the
original complex and complete formation of the new species
and free ligand (Figure 10). A similar timescale is observed
for the two processes in which both complexes release
ꢀ20% of the bis(phenolato) ligand as free ligand H2L and
the final complicated spectrum clearly reveals substantial
Figure 8. 1H NMR spectra of [Ti(L1)
ACTHNUGTRENNUG(OiPr)2] at RT a) in [D8]THF, b) im-
mediately following the addition of D2O, and c) 3.5 h following the addi-
tion of D2O.
2408
www.chemeurj.org
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 2403 – 2415

TiIV Complexes
FULL PAPER
Figure 12. A plot of the integration of selected signals in the 1H NMR
spectrum of [Ti(L5)
ACHTUNGTRENNUNG
Figure 10. A plot of the integration of selected signals in the 1H NMR
spectrum of [Ti(L1)
ACHTUNGTRENNUNG(OiPr)2] versus time following addition of D2O to a
solution of the complex in [D8]THF at a) RT and b) 378C; [a] The inte-
gration values for this particular plot of the “new species” were not cali-
brated like the rest[41b] due to the complexity of the spectrum and difficul-
ty in assigning particular proton identity.
Figure 13. 1H NMR spectrum of a) [Ti(L1)(OiPr)2] and b) [Ti(L3)
ACHUTGTNERNNUG ACHTUNGTRENNUNG(OiPr)2]
in [D8]THF 14 h following the addition of D2O at 378C.
less than that of the starting complex, it is clear that the
signal of the new species selected for the plot corresponds
to a smaller number of protons, a fact that is also supported
by its low symmetry as apparent by its multiple signals. As
Figure 11. A plot of the integration of selected signals in the 1H NMR
spectrum of [Ti(L3)
ACHTUNGTRENNUNG(OiPr)2] versus time following the addition of D2O to
a solution of the complex in [D8]THF at RT.
opposed to [Ti(Ln)
ducted with [Ti(L3)
ACTHNUGNRTE(NUG OiPr)2] at 378C supported complete
ACHTUNGTNER(NUNG OiPr)2] (n=1 and 2), an experiment con-
formation of a new compound (Figure 13). As the sum of
the final integration of the free ligand and new species is
ligand release to give only H2L3 and isopropanol
(Figure 13).
Chem. Eur. J. 2009, 15, 2403 – 2415
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
2409

E. Y. Tshuva et al.
To summarize, similar general hydrolytic behavior is ob-
served for the two similarly active complexes [Ti(Ln)
(OiPr)2]
AHCTUNGTRENNUNG
(n=1 and 2), namely, significant amount of new species for-
mation relative to free bis(phenolato) ligand, with slower
hydrolysis for the bulkier complex (Table 5). In contrast, the
Table 5. t = values for hydrolysis[a] of [Ti(Ln)
(OiPr)2] (n=1–5) based on
1
2
NMR measurements.
1
1
t =
2
Complex
t =
Complex
2
[Ti(L1)
[Ti(L2)
[Ti(L3)
C
5 h[b]
[Ti(L4)
[Ti(L5)
C
3 h
5 min
31 h[b]
10 h[b]
À
[a] Hydrolysis rate calculated based on the release of Ti OiPr groups to
give isopropanol. [b] Calculated based on trend-line.
inactive complex [Ti(L3)
ACTHNUTRGNEU(GN OiPr)2] with bulkier ortho-substitu-
ents leads to complete free bis(phenolato) ligand release si-
multaneously with isopropanol release with no new species
formation, while the inactive complex [Ti(L5)
ACHTUNGTRENNUNG
Figure 14. An ORTEP drawing of [Ti3(L1)
ellipsoids; H atoms and diethylether solvent were omitted for clarity.
3ACHTNUGTREN(NNUG m2-O)3] with 50% probability
ACHTUNGTRENNUNG
Table 6. Selected bond lengths [ꢁ] and angles [8] for [Ti3(L1)
3ACHTUNGTRENNUNG
(m2-O)3].
ACHTUNGTRENNUNG
groups exhibit hydrolysis rates of the same order of magni-
tude despite the difference in cytotoxic activity, and [Ti(Ln)-
ACHTUNGTRENNUNG(OiPr)2] (n=3 and 5) , which are similarly inactive biologi-
À
À
O(1) Ti(2)
1.776(3)
1.925(3)
1.805(3)
1.855(3)
1.773(3)
1.911(3)
1.877(3)
1.921(3)
1.920(3)
1.914(3)
O(8) Ti(3)
1.864(3)
1.905(3)
À
À
O(1) Ti(1)
O(9) Ti(3)
À
O(2) Ti(3)
À
O(2) Ti(2)
cally, demonstrate significantly different rates of hydrolysis
(Table 5). It is also quite surprising that in all processes in
which a bis(phenolato) ligand is released, its hydrolysis rate
is similar to that of the isopropoxo groups, which are hydro-
À
À
O(3) Ti(1)
N(1) Ti(1)
2.369(4)
2.293(4)
2.414(4)
2.361(4)
2.387(4)
2.284(4)
À
À
O(3) Ti(3)
N(2) Ti(1)
À
À
O(4) Ti(1)
N(3) Ti(2)
À
À
O(5) Ti(1)
N(4) Ti(2)
ACTHNUTRGNEUNG
lyzed quite slowly for [Ti(Ln)(OiPr)2] (n=1–4).[42–46] Overall,
À
À
O(6) Ti(2)
N(5) Ti(3)
À
À
O(7) Ti(2)
N(6) Ti(3)
these observations are consistent with those obtained by the
UV/Vis measurements regarding formation of a new species
in substantial amounts for the biologically active complexes
O(4)-Ti(1)-O(3)
O(4)-Ti(1)-O(1)
O(3)-Ti(1)-O(1)
O(4)-Ti(1)-O(5)
O(3)-Ti(1)-O(5)
O(1)-Ti(1)-O(5)
O(4)-Ti(1)-N(1)
O(3)-Ti(1)-N(1)
O(1)-Ti(1)-N(1)
O(5)-Ti(1)-N(1)
O(4)-Ti(1)-N(2)
O(3)-Ti(1)-N(2)
O(1)-Ti(1)-N(2)
O(5)-Ti(1)-N(2)
N(1)-Ti(1)-N(2)
O(7)-Ti(2)-O(6)
O(7)-Ti(2)-O(1)
O(6)-Ti(2)-O(1)
104.07(12)
96.24(13)
95.37(13)
92.86(13)
95.88(14)
163.37(13)
80.92(13)
175.01(13)
83.90(12)
83.84(14)
157.54(13)
98.04(13)
85.44(12)
80.90(13)
76.99(13)
162.29(13)
98.18(13)
94.52(13)
O(1)-Ti(2)-N(3)
O(2)-Ti(2)-N(3)
O(7)-Ti(2)-N(4)
O(6)-Ti(2)-N(4)
O(1)-Ti(2)-N(4)
O(2)-Ti(2)-N(4)
N(3)-Ti(2)-N(4)
O(8)-Ti(3)-O(3)
O(8)-Ti(3)-O(9)
O(3)-Ti(3)-O(9)
O(8)-Ti(3)-O(2)
O(3)-Ti(3)-O(2)
O(9)-Ti(3)-O(2)
O(8)-Ti(3)-N(5)
O(3)-Ti(3)-N(5)
O(9)-Ti(3)-N(5)
O(2)-Ti(3)-N(5)
O(8)-Ti(3)-N(6)
O(3)-Ti(3)-N(6)
O(9)-Ti(3)-N(6)
O(2)-Ti(3)-N(6)
N(6)-Ti(3)-N(5)
O(1)-Ti(2)-N(3)
O(2)-Ti(2)-N(3)
167.33(13)
88.45(13)
78.02(12)
88.81(12)
94.20(13)
160.50(14)
74.31(13)
93.85(13)
93.02(14)
164.30(13)
107.75(13)
94.06(13)
97.26(15)
83.84(13)
81.94(13)
84.78(15)
168.04(13)
160.07(14)
[Ti(Ln)
ACHTUNGTRENNUNG
AHCTUNGTERN(GNUN OiPr)2]
was reacted with specific equivalents of water in an attempt
to replace the isopropoxo groups and to produce O-bridged
clusters with a bound diamine bis(phenolato) ligand, and
perhaps gain structural information about the new species
1
observed to form in the H NMR measurements. Addition
of one, two, or ten equivalents of water to a solution of
[Ti(L1)
ACHTUNGTRENNUNG(OiPr)2] in THF and stirring at room temperature for
three days gave a significant amount of single crystals that,
somewhat surprisingly,[42–46] turned out to be the starting
complex [Ti(L1)
ACHTUNGTRENNUNG(OiPr)2] (Figure 1, top). However, whereas
increasing the amount of water to 100 equivalents gave simi-
lar results when the reaction was mixed for only 1.5 h, differ-
ent single crystals suitable for X-ray crystallography were
obtained in 95% yield when the reaction was allowed to stir
with 50 equivalents of water for three days. Similar crystals
were also obtained when the amount of water varied be-
tween 50 and 100000 equivalents. The structure of the prod-
uct in all of these cases is presented in Figure 14 and select-
ed bond lengths and angles are summarized in Table 6.
O(7)-Ti(2)-O(2)
O(6)-Ti(2)-O(2)
O(1)-Ti(2)-O(2)
O(7)-Ti(2)-N(3)
O(6)-Ti(2)-N(3)
91.65(13)
97.31(13)
103.70(13)
84.83(14)
80.18(14)
85.52(13)
83.22(14)
92.16(14)
76.34(14)
167.33(13)
The structure contains a trinuclear complex with three m2-
O ligands. Interestingly, two out of the three TiIV centers
2410
www.chemeurj.org
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 2403 – 2415

TiIV Complexes
FULL PAPER
(Ti(1) and Ti(3)) feature cis-phenolato groups on the tetra-
dentate ligand rather than a trans configuration as in the
starting complex, and similar to the catecholato substituted
free catechol and free ligand H2L1 of around 10%, with no
new signals that could be clearly detected (Figure 16). Since
following the hypothetical hydrolysis of the catecholato
ACHTUNGTRENNUNG
product [Ti(L1)(O2Ph)][39,47] (Figure 1, bottom), suggesting a
similar mechanism for ligand replacement. A partial draw-
ing that includes such a TiIV center and its bound ligands is
presented in Figure 15. The Ti···Ti distances of the trinuclear
Figure 15. A partial ORTEP drawing of [Ti3(L1)
ability ellipsoids, which includes a single TiIV center.
3ACHTUNGTREN(UNGN m2-O)3] with 50% prob-
Figure 16. A plot of the integration of selected signals in the 1H NMR
spectrum of [Ti(L1)
ACTHUNTRGNEUNG(O2Ph)] versus time following the addition of D2O to
a solution of the complex in [D8]THF at RT.
À
core are in the range of 3.4–3.5 ꢁ, with Ti m-O bond lengths
of 1.77–1.93 ꢁ and Ti-m-O-Ti and m-O-Ti-m-O angles of
133.6–145.38 and 94.1–103.88, respectively. Notably, the TiIV
center of the trans-phenolato group exhibits a substantially
wider m-O-Ti-m-O angle than that of the other two, as well
ligand the same L1 ligand and its particular steric require-
ments remain as in [Ti(L1)
ACTHNUTRGEN(UNG OiPr)2], which is able to form a
trinuclear complex, it is plausible that the strong binding of
the catecholato ligand and perhaps also the different sym-
metry of the original complex prohibit this interaction, as
the complex symmetry was previously observed to have a
strong influence on activity.[33] Interestingly, this more inert
complex also exhibits reduced cytotoxicity against both HT-
À
À
as somewhat longer Ti O and Ti N bonds. Altogether, this
structure exhibits a low symmetry of C1, which explains the
complicated NMR spectra obtained for the new species
formed within several hours following water addition, and in
particular, the integration ratio of the ligand signals (see
above). Interestingly, the smallest distance between the two
aromatic carbon atoms ortho to the phenolato oxygen atoms
of two different bis(phenolato) ligands is 3.7 ꢁ, with the dis-
tance between their corresponding hydrogen atoms being
3.3 ꢁ. It is, therefore, logical that the ortho-methyl substitut-
ed analogue might be able to give an analogous structure
through longer reaction times (Table 5), while an ortho-tert-
butyl substituted analogue of this particular constrained
structure is inconceivable. The 1H NMR spectrum and
LMCT absorption of the trinuclear complex are similar to
those of the new species formed as described above. There-
fore, if steric crowding allows, a new polynuclear species
may form upon replacement of the isopropoxo groups; how-
ever, this process is not rapid for this family of com-
plexes.[42–46]
29 and OVCAR-1 cells relative to [Ti(L1)
ACHTUNGTRENNUNG(OiPr)2]
(Figure 17).[23] This observation supports the notion that the
ligand steric requirements allow for some activity, yet some
release of the labile groups needs to take place in order for
Effect of ligand lability in [Ti(L1)
two labile OiPr groups with a single bidentate ligand in
[Ti(L1)
(O2Ph)] gave a complex of enhanced hydrolytic sta-
ACHTUNGTREN(UNGN O2Ph)]: Replacing the
ACHTUNGTRENNUNG
bility as expected. Since the LMCT band of the bis(phenola-
to) ligand overlaps with that of the catecholato unit, analysis
by UV/Vis is problematic. The 1H NMR measured over
Figure 17. Dependence of HT-29 cell viability after 3 days of incubation
on concentration of [Ti(L1)(OiPr)2], [Ti(L1)(O2Ph)], and [Ti3(L1)
ACHTUNGRTENUNNG ACHUTNTGRNNEUG CHTUNGTRENNUNG(m2-O)3]
3A
[Ti3(L1)
: 3ACHTUNGTRENNUNG(m2-O)3];
time upon D2O addition to a solution of [Ti(L1)
ACHTUNGTRENNUNG
^
administered presented in logarithmic scale.
1
1
~
&
: [Ti(L )ACTHNUGRTENN(UG O2Ph)]; : [Ti(L )ACHTUTGNREN(NUGN OiPr)2].
Chem. Eur. J. 2009, 15, 2403 – 2415
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
2411

E. Y. Tshuva et al.
the complex to demonstrate activity. Perhaps, following this
release, the symmetry of the original complex, namely the
initial geometry of the ligand binding, is of lesser impor-
tance, as rearrangement could occur. This behavior is in gen-
eral different from that observed with titanocene dichloride
and its analogues, in which replacement of two chloride li-
gands with a bidentate oxalate significantly enhanced activi-
ty.[48]
ACHTUNGTERN(NGNU OiPr)2] (n=1 and 4) despite their different cytotoxicity;
[Ti(L2)
[Ti(L1)
AHCTUNGTRENNUNG
ACHTUNGTRENNUNG
sis despite the similar cytotoxicity of the two, presumably
due to slower polynuclear complex formation requiring the
proximity of two metal centers, while [Ti(L3)
ACTHUNTRGNE(UGN OiPr)2] does
not form such a species (Table 5). Additionally, increasing
steric bulk directly on the donor to the metal has an im-
mensely negative effect on the cytotoxicity. As larger sub-
stituents should inhibit hydrolysis for complexes with strong
chelating ligand binding, the substantially more rapid hy-
What is the active species? As the active complexes [Ti(Ln)-
ACHTUNGTRENNUNG(OiPr)2] (n=1, 2, and 4) showed indications for rather sub-
stantial formation of an O-bridged bis(phenolato) polynuc-
lear species upon water addition, while the inactive complex
drolysis observed for [Ti(L5)
binding in solution despite the similar Ti N bonds observed
in the solid-state structures of [Ti(Ln)
(OiPr)2] (n=1 and 5),
and, thus, the reduced cytotoxicity of [Ti(L5)
(OiPr)2] pre-
ACHTUNGERTN(NUNG OiPr)2] points to weaker ligand
À
[Ti(L3)
N
ACHTUNGTRENNUNG
lease, we were interested in exploring whether the polynuc-
lear complexes might be the actual active species as suggest-
ed for budotitane derivatives,[18–20,49] despite the apparent
formation of smaller amounts of a polynuclear cluster from
ACHTUNGTRENNUNG
sumably results from rapid formation of unreactive aggre-
gates, which emphasizes the importance of a strongly bound
chelating ligand as supported by the inactivity of particularly
labile complexes.[23] Thus, we may generalize and say that
mostly bis(phenolato) ligand release (whether it comes from
weaker ligand binding, and thus a hydrolytically instable
complex, or from large steric groups) vs. substantial poly-
nuclear complex formation upon water addition, is more sig-
nificant in determining biological activity than the particular
rate of hydrolysis. This is particularly demonstrated by the
the inactive [Ti(L5)(OiPr)2] as well. [Ti3(L1)
ACHTUNGTRENNUNG 3AHCTUNGTREN(NGUN m2-O)3], which
is itself stable for days in the presence of water, was found
to be inactive against the colon and ovarian cells measured
(Figure 17). Thus, the well-defined cluster is not the active
species formed through the hydrolysis of [Ti(L1)
ACHTUNGTRENNUNG
ACHTUNGTRENNUNG
1
itself, then perhaps an intermediate in its formation might
lead to activity, and, in general, the characteristics that allow
its formation in terms of steric requirement are essential, al-
though not suffice, for obtaining cytotoxic activity for this
family of compounds.
immense difference in t = values of hydrolysis for the two
2
similarly inactive complexes, [Ti(Ln)
ACHTUNGTRENNUNG
AHCTUNGTRENNUNG
and 2). Thus, a combination of strong chelating ligand bind-
ing and reduced steric bulk all around is required for activi-
ty. Overall, it is evident that careful design of particular li-
gands is essential in obtaining cytotoxicity of TiIV complexes
as the ligand features have an immense influence on the hy-
drolysis and biological activity, which are not unrelated,
Conclusion
In this work, we presented several octahedral TiIV com-
plexes of tetradentate diamine bis(phenolato) ligands and
explored the influence of steric effects and lability of the li-
gands on their cytotoxicity and hydrolytic behavior. Bulky
groups have a negative effect on cytotoxicity, not merely
when located in close proximity to the metal site, but also
when located on the ligand periphery, judging from the simi-
while TiO2 and labile [Ti
ACHUTNGTREN(GUN OiPr)4] and [TiCl4ACHTUNGTREN(NGUN thf)2] are all in-
active.[23]
As discussed, the ability to form polynuclear species
through release of isopropoxo ligands upon water addition
might be an important factor in obtaining cytotoxic agents
from this family of compounds. This process becomes signifi-
cantly disfavored for the catecholato complex, which shows
not only enhanced hydrolytic stability, but also reduced cy-
totoxicity relative to its bis(isopropoxo) counterpart with
the same tetradentate ligand. Still, one should not forget
that the general geometry and symmetry of the catecholato
complex is different from those of its bis(isopropoxo) coun-
terpart, a parameter that itself may have a crucial influence
on activity as we previously reported.[33] Nevertheless, it is
plausible that the cytotoxic activity for these particular com-
plexes[23] requires release of some labile ligands, especially
considering the activity mechanism of cisplatin that involves
DNA binding through the positions of the chloro groups fol-
lowing their hydrolysis. Thus, a combination of some lability
of the monodentate groups and relative inertness of the bi-
s(phenolato) ligand (see above) are essential. In addition,
since the trinuclear complex isolated from the hydrolysis re-
action of an active complex is itself inactive, the ability to
lar cytotoxic activity of [Ti(L1)
and lower activity of [Ti(L4)
(OiPr)2]. As the ability to form
a polynuclear species upon water addition correlates with
notable cytotoxic activity for [Ti(Ln)
(OiPr)2] (n=1–4), we
(OiPr)2] and [Ti(L2)
ACHUTGTNERNNUG AHCUTNGTREUNGN(N OiPr)2]
AHCTUNGTRENNUNG
AHCTUNGTRENNUNG
suspect that as large substituents near the metal center
block the approach of two ligand-bound metal centers re-
quired to form a cluster, they may also block other biologi-
cal interactions valuable for activity. More distant bulky sub-
stituent may have a similar influence; however, they may
also point to some involvement of DNA intercalation in the
biological activity, for which high ligand planarity is re-
quired. However, these observations may also reflect the
effect of other parameters, such as solubility and/or cell pen-
etration rate. In contrast, the near-metal-site steric effects
represent the biggest influence on the hydrolysis rate: rather
similar t = values for hydrolysis are observed for [Ti(Ln)-
1
2
2412
www.chemeurj.org
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 2403 – 2415

TiIV Complexes
FULL PAPER
([Ti(L5)
ACTHNUTRGNEUNG(OiPr)2]) in pure water are still unreactive under biological condi-
tions, we conclude that the reactivity measured for their complexes in
vitro does not result from hydrolyzed ligand.
form it may point to an original complex of steric require-
ments suitable for its interaction with the biological target
that leads to its activity, an interaction that must occur quite
rapidly considering the deactivation of the complex after
long periods in medium. Additional UV/Vis experiments in-
volving the addition of biological medium (RPMI-1640)
Synthesis of H2L4: H2L4 was synthesized by mixing 4-tert-butylphenol
(20.0 mmol) with paraformaldehyde (20.0 mmol) and N,N’-dimethylethy-
lenediamine (10.0 mmol) in methanol and heating to 508C for 3 days.
The crude product was cooled, filtered, and washed with cold methanol
to yield H2L4 in 37%. 1H NMR (400 MHz, CDCl3): d=7.2 (dd, J=8.6,
2.4 Hz, 2H; Ar), 6.9 (s, 2H; Ar), 6.8 (d, J=8.6 Hz, 2H; Ar), 3.7 (s, 4H;
CH2), 2.7 (s, 4H; CH2), 2.3 (s, 6H; NMe), 1.3 ppm (s, 18H; tBu);
13C NMR (400 MHz; CDCl3): d=155.3, 141.8, 125.6, 125.2, 120.8, 115.6,
62.2, 54.3, 41.8, 33.9, 31.6 ppm; elemental analysis calcd (%) for
C26H40N2O2 (412.6): C 75.68, H 9.77, N 6.79; found: C 75.93, H 9.95, N
6.94.
Synthesis of H2L5: H2L5 was synthesized by heating 3,4-dimethylphenol
(5.0 mmol) with formaldehyde (10.0 mmol) and N,N’-diethylethylenedia-
mine (2.5 mmol) to reflux in methanol for 5 h. The crude product was
cooled, filtered, and washed with cold methanol to yield H2L5 in 67%.
1H NMR (300 MHz, CDCl3): d=6.7 (s, 2H; Ar), 6.6 (s, 2H; Ar), 3.7 (s,
4H; CH2), 2.7 (s, 4H; CH2), 2.6 (q, J=7.06 Hz, 4H; CH2CH3), 2.2 (s,
6H; ArCH3), 2.1 (s, 6H; ArCH3), 1.0 ppm (t, J=7.06 Hz, 6H; CH2CH3);
13C NMR (400 MHz; CDCl3): d=155.6, 136.9, 129.5, 126.8, 118.8, 117.3,
57.5, 50.4, 47.4, 19.6, 18.7, 11.0 ppm; elemental analysis calcd (%) for
C24H36N2O2 (384.6): C 74.96, H 9.44, N 7.28; found: C 74.70, H 9.56, N
7.09.
rather than pure water to [Ti(Ln)
ACHTUNGTERN(NNUG OiPr)2] (n=1–3) and
[Ti3(L1)
CHTUNGTRENNUNG
shift in lmax for all of these complexes within several hours.
Thus, it seems as if the active species is only rapidly formed
in the presence of cells where steric crowding and electronic
À
features allow, and once Ti O aggregates are formed and/or
ligands are completely released, the biological reactivity is
lost. Therefore, the distinctive characteristics of the family
of complexes presented herein regarding mediocre rate of
hydrolysis of their OiPr groups is of particular merit for bio-
logical applications.
Experimental Section
Synthesis of [Ti(L4)
(OiPr)2]: [Ti(L4)
ACHUTNGTRENNUG CAHTUNGTREN(NGUN OiPr)2] was synthesized by treating
Ligands and their bis(isopropoxo) TiIV complexes were synthesized ac-
[Ti
AHCTUNGTRENNUNG
cording to published procedures.[35–38] Data on [Ti(Ln)
ACHTNUTRGNE(UNG OiPr)2] (n=1–3)
RT under a nitrogen atmosphere to give a yellow product in a quantita-
tive yield. Yellow single crystals were obtained from diethyl ether by
slow evaporation at RT. 1H NMR (400 MHz, CDCl3): d=7.2 (ddd, J=
8.4, 2.6, 0.6 Hz, 2H; Ar), 6.9 (d, J=2.4 Hz, 2H; Ar), 6.6 (d, J=8.3 Hz,
2H; Ar), 5.0 (sept, J=6.1 Hz, 2H; CHCH3), 4.7 (d, J=13.2 Hz, 2H;
CH2), 3.1 (d, J=13.4 Hz, 2H; CH2), 3.0 (d J=9.3, Hz, 2H; CH2), 2.5 (s,
6H; NMe), 1.8 (d, J=9.3 Hz, 2H; CH2), 1.3 (s, 18H; tBu), 1.3 (d, J=
6.2 Hz, 6H; CHCH3), 1.3 ppm (d, J=6.0 Hz, 6H; CHCH3); 13C NMR
(400 MHz, CDCl3): d=159.6, 140.1, 126.0, 125.6, 123.8, 116.6, 77.6, 64.9,
51.8, 47.3, 33.9, 31.7, 26.0, 25.7 ppm; UV/Vis (THF): lmax (e)=328 nm
(17210mÀ1 cmÀ1); elemental analysis calcd (%) for C32H52N2O4Ti (576.3):
C 66.65, H 9.09, N 4.86; found: C 66.93, H 9.31, N 5.03.
can be found elsewhere.[33,36,37,50] Paraformaldehyde (95%) was purchased
from Fluka Chemica and formaldehyde 30–38% assay was purchased
from Bio Lab Ltd. and used without further purification. N,N’-Diethyle-
thylenediamine (96%) was purchased from Alfa-Aesar a Johnson Mat-
they Company and used without further purification. Pyrocatechol
(99%), 4-tert-butylphenol (99%), 3,4-dimethylphenol (99%), and N,N’-
dimethylethylenediamine (99%) were purchased from Sigma-Aldrich
Chemical Company and used without further purification. Titanium tet-
ra(isopropoxide) (97%) was purchased from Aldrich Chemical Company.
[TiCl2Cp2] for reference measurements was purchased from Arapahoe
Chemicals. All solvents were distilled from K or K/benzophenone under
nitrogen. All experiments requiring dry atmosphere were performed in a
M. Braun dry-box or under nitrogen atmosphere using Schlenk tech-
niques. NMR data were recorded using an AMX-300, 400 or 500 MHz
Bruker spectrometer. CDCl3 (99.8%) and D2O (99.9%) were purchased
from Sigma-Aldrich Chemical Company and used without further purifi-
cation. [D8]THF (99.5%) was purchased from D-Chem and used without
further purification. Hydrolysis experiments followed by NMR spectros-
copy were conducted with a mixture of 90% [D8]THF and 10% D2O, in
which the analysis was performed every 4–10 minutes. UV/Vis spectra
were recorded on a Jasco V-530 spectrophotometer for which solutions (
ꢀ0.05 mm) of the complexes in 90% THF and 10% water were used for
the hydrolysis studies. Experiments involving the addition of biological
medium were performed with RPMI-1640 medium purchased from
Sigma without glutamic acid and without phenol red. Stopped flow meas-
urements were performed on a HiTech Scientific diode-array instrument
under similar conditions. Hydrolysis reactions for identifying O-bridged
polynuclear complexes were performed on solutions of the complex
(1.8 mmol) in THF by adding 1, 2, 10, 50, 100, 1000, 10000 or
100000 equivalents of water. X-ray diffraction data were obtained with
Bruker Smart Apex diffractometer, running the SMART software pack-
age.[51] After collection, the raw data frames were integrated by the
Crystal data for [Ti(L4)
a=9.6622(7), b=27.033(2), c=13.4165(9) ꢁ, b=106.378(1)8, V=
3362.2(4) ꢁ3, T=223(1) K, space group P21/c, Z=4,
(MoKa)=
0.289 mmÀ1
34755 reflections measured, 6593 unique (Rint =0.0461);
R(F2) for [I>2s(I)]=0.0839, Rw for [I>2s(I)]=0.1860.
Synthesis of [Ti(L5)(OiPr)2]: [Ti(L5)
(OiPr)2] was synthesized similarly as
ACTHUGNNERTN(NUG OiPr)2]: C32H52N2O4Ti, M=576.66, monoclinic,
mACHTUNGTRENNUNG
,
A
ACHTUNGTRENNUNG
a yellow solid in quantitative yield by treating [TiACTHNURGTNEUNG(OiPr)4] (0.2 mmol)
with H2L5 (0.2 mmol) in THF (5 mL) for 2 h at RT. Yellow single crystals
1
were obtained from diethyl ether at À358C. H NMR (400 MHz, CDCl3):
d=6.7 (s, 2H; Ar), 6.5 (s, 2H; Ar), 4.9 (m, J=6.1 Hz, 2H; CHACHTUNGTRENNUNG(CH3)2),
4.5 (d, J=13.3 Hz, 2H; CH2), 3.2 (dq, J=13.7, 7.0 Hz, 2H; CH2), 3.1 (d,
J=13.3 Hz, 2H; CH2), 2.9 (d, J=10.1 Hz, 2H; CH2), 2.4 (dq, J=13.6,
6.9 Hz, 2H; CH2), 2.2 (s, 6H; CH3), 2.1 (s, 6H; CH3), 1.8 (d, J=10.2 Hz,
2H; CH2), 1.3 (t, J=7.2 Hz, 6H; CH2CH3), 1.2 (d, J=6.4 Hz, 6H; CH-
(CH3)2); 13C NMR (400 MHz,
ACHUTNGRENN(UG CH3)2), 1.1 ppm (d, J=6.1 Hz, 6H; CHCAHTUNGTRENNUGN
CDCl3): d=159.9, 136.8, 130.6, 125.0, 121.7, 118.2, 77.5, 58.2, 51.3, 48.2,
26.0, 25.9, 19.6, 18.7, 8.81 ppm; UV/Vis (THF): lmax (e)=320 nm
(20790mÀ1 cmÀ1); elemental analysis calcd (%) for C30H48N2O4Ti (548.6):
C 65.68, H 8.82, N 5.11; found: C 65.64, H 9.03, N 4.94.
Crystal data for [Ti(L5)
a=15.017(1), b=13.260(1), c=15.157(1) ꢁ, b=100.960(18), V=
2963.0(4) ꢁ3, T=173(1) K, space group P21/n, Z=4,
(MoKa)=
0.325 mmÀ1
30243 reflections measured, 5827 unique (Rint =0.0485);
R(F2) for [I>2s(I)]=0.0712, Rw for [I>2s(I)]=0.1498.
Synthesis of [Ti(L1)(O2Ph)]: [Ti(L1)
(O2Ph)] was synthesized quantitative-
ly by treating [Ti(L1)
(OiPr)2] (0.2 mmol) with pyrocatechol (0.2 mmol) in
ACTHUGNNERTN(NUG OiPr)2]: C30H48N2O4Ti, M=548.60, monoclinic,
SAINT software package.[52] The structures were solved and refined
[53]
mACHTUNGTRENNUNG
using the SHELXTL software package.
Elemental analyses were per-
,
formed in the microanalytical laboratory in our institute. Cytotoxicity
was measured on HT-29 colon and OVCAR-1 ovarian cells obtained
from ATCC using the methylthiazolyldiphenyl-tetrazolium bromide
(MTT) assay as previously described.[33] Most complexes demonstrate in-
complete solubility when the highest concentration is applied. Two of the
free ligands demonstrate some cytotoxic activity (H2L4, H2L5); however,
since complexes that rapidly hydrolyze to release the free ligands
A
ACHTUNGTRENNUNG
AHCTUNGTRENNUNG
THF for 4 h at RT and under a nitrogen atmosphere to give a dark red
solution. Following evaporation of the solvent, red single crystals were
1
obtained from diethylether at RT. H NMR (400 MHz, CDCl3): d=6.9 (s,
Chem. Eur. J. 2009, 15, 2403 – 2415
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
2413

E. Y. Tshuva et al.
1H; Ar), 6.8 (s, 1H; Ar), 6.7 (s, 1H; Ar), 6.6 (dt, J=7.6, 1.6 Hz, 1H; Ar-
cat), 6.5 (dt, J=7.6, 1.6 Hz, 1H; Ar-cat; cat=catecholato), 6.5 (s, 1H;
Ar), 6.4 (dd, J=7.6, 1.6 Hz, 1H; Ar-cat), 6.2 (dd, J=7.6, 1.6 Hz, 1H; Ar-
cat), 5.2 (d, J=13.1 Hz, 1H; ArCH2N), 4.7 (d, J=13.5 Hz, 1H;
ArCH2N), 3.7 (dt, J=13.4, 3.4 Hz, 1H; NCH2CH2), 3.4 (d, J=13.5 Hz,
1H; ArCH2N), 3.2 (d, J=13.2 Hz, 1H; ArCH2N), 3.1 (dt, J=13.6,
2.8 Hz, 1H; NCH2CH2), 2.7 (s, 3H; NMe), 2.5 (dd, J=12.8, 1.8 Hz, 1H;
NCH2CH2), 2.2 (m, 1H; NCH2CH2), 2.2 (s, 3H; NMe), 2.2 (s, 3H;
ArMe), 2.2 (s, 3H; ArMe), 2.2 (s, 3H; ArMe), 2.1 ppm (s, 3H; ArMe);
13C NMR (400 MHz, CDCl3): d=194.0, 160.2, 159.6, 158.6, 158.5, 138.0,
137.2, 130.6, 129.6, 128.7, 127.4, 122.8, 120.6, 120.1, 117.7, 117.6, 112.5,
111.5, 63.9, 63.4, 57.6, 52.0, 49.0, 42.9, 19.6, 19.5, 18.9, 18.8 ppm; UV/Vis
(THF): lmax (e)=344 nm (14610mÀ1 cmÀ1); elemental analysis calcd (%)
for C28H34N2O4Ti (510.2): C 65.88, H 6.71, N 5.49; found: C 65.59, H 6.72,
N 5.46.
[16] G. Kelter, N. J. Sweeney, K. Strohfeldt, H.-H. Fiebig, M. Tacke,
Anti-Cancer Drugs 2005, 16, 1091–1098.
[17] B. K. Keppler, C. Friesen, H. G. Moritz, H. Vongerichten, E. Vogel,
Struct. Bonding 1991, 78, 97–127.
[18] F. Caruso, M. Rossi, J. Tanski, R. Sartori, R. Sariego, S. Moya, S.
Diez, E. Navarrete, A. Cingolani, F. Marchetti, C. Pettinari, J. Med.
Chem. 2000, 43, 3665–3670.
[19] F. Caruso, M. Rossi, Mini-Rev. Med. Chem. 2004, 4, 49–60.
[20] F. Caruso, M. Rossi, C. Opazo, C. Pettinari, Bioinorg. Chem. Appl.
2005, 3, 317–329.
[21] J. H. Toney, T. J. Marks, J. Am. Chem. Soc. 1985, 107, 947–953.
[22] P. Yang, M. Guo, Coord. Chem. Rev. 1999, 185, 189–211.
[23] M. Shavit, D. Peri, A. Melman, E. Y. Tshuva, J. Biol. Inorg. Chem.
2007, 12, 825–830.
[24] M. Ravera, C. Cassino, E. Monti, M. Gariboldi, D. Osella, J. Inorg.
Biochem. 2005, 99, 2264–2269.
[25] F. Caruso, L. Massa, A. Gindulyte, C. Pettinari, F. Marchetti, R. Pet-
tinari, M. Ricciutelli, J. Costamagna, J. C. Canales, J. Tanski, M.
Rossi, Eur. J. Inorg. Chem. 2003, 3221–3232.
[26] S. H. Sun, H. Li, R. A. Weir, P. J. Sadler, Angew. Chem. 1998, 110,
1622–1625; Angew. Chem. Int. Ed. 1998, 37, 1577–1579.
[27] M. Guo, H. Sun, H. J. McArdle, L. Gambling, P. J. Sadler, Biochem-
istry 2000, 39, 10023–10033.
[28] M. Guo, H. Sun, S. Bihari, J. A. Parkinson, R. O. Gould, S. Parsons,
P. J. Sadler, Inorg. Chem. 2000, 39, 206–215.
[29] L. M. Gao, R. Hernandez, J. Matta, E. Melendez, J. Biol. Inorg.
Chem. 2007, 12, 959–967.
[30] A. D. Tinoco, C. D. Incarvito, A. M. Valentine, J. Am. Chem. Soc.
2007, 129, 3444–3454.
[31] A. D. Tinoco, A. M. Valentine, J. Am. Chem. Soc. 2005, 127, 11218–
11219.
[32] A. D. Tinoco, E. V. Eames, A. M. Valentine, J. Am. Chem. Soc.
2008, 130, 2262–2270.
[33] M. Shavit, D. Peri, C. M. Manna, J. S. Alexander, E. Y. Tshuva, J.
Am. Chem. Soc. 2007, 129, 12098–12099.
[34] E. Dubler, R. Buschmann, H. W. Schmalle, J. Inorg. Biochem. 2003,
95, 97–104.
[35] S. Gendler, S. Segal, I. Goldberg, Z. Goldschmidt, M. Kol, Inorg.
Chem. 2006, 45, 4783–4790.
Crystal data for [Ti(L1)
a=7.9048(8), b=19.278(2), c=16.699(2) ꢁ, b=95.756(2)8, V=
2531.9(4) ꢁ3, T=295(1) K, space group P21/c, Z=4,
(MoKa)=
0.375 mmÀ1
26519 reflections measured, 4979 unique (Rint =0.0847);
R(F2) for [I>2s(I)]=0.0816, Rw for [I>2s(I)]=0.1451.
3A 3A(m2-O)3] was synthesized by treat-
Synthesis of [Ti3(L1) (m2-O)3]: [Ti3(L1)
ing (OiPr)2] (0.2 mmol) in THF (30 mL) with
solution of [Ti(L1)
ACHTUNGTRENNUNG(O2Ph)]: C28H34N2O4Ti, M=510.47, monoclinic,
mACHTUNGTRENNUNG
,
C
CHTUNGTRENNUNG
a
ACHTUNGTRENNUNG
50 equivalents of water in THF (10 mL) at RT under a nitrogen atmos-
phere for 3 days. Recrystallization from diethylether gave yellow single
crystals in 95% yield. UV/Vis (THF): lmax (e)=336 nm (27990mÀ1 cmÀ1);
elemental analysis calcd (%) for C66H90N6O9Ti3 (1255.2): C 63.16,H 7.23,
N 6.70; found: C 63.19, H 7.53, N 6.50.
Crystal data for [Ti3(L1)
3ACTHGNUNERTUN(GN m2-O)3]: C66H90N6O9Ti3·C4H10O, M=1329.26,
monoclinic, a=13.4535(9), b=32.001(2), c=16.669(1) ꢁ, b=104.678(1)8,
V=7775.2(9) ꢁ3, T=173(1) K, space group P21/c, Z=4,
mACHTNGUTRENNU(G MoKa)=
0.354 mmÀ1, 80097 reflections measured, 15239 unique (Rint =0.0418).
R(F2) for [I>2s(I)]=0.0991, Rw for [I>2s(I)]=0.2346.
CCDC 692997 ([Ti(L4)
([Ti(L1)(O2Ph)]), and 693000 ([Ti3(L1)
CHTUNGTRENNUNG
N
ACHTUNGTERN(NGNU OiPr)2]), 692999
U
tary crystallographic data for this paper. These data can be obtained free
of charge from The Cambridge Crystallographic Data Centre via
www.ccdc.cam.ac.uk/data_request/cif.
[36] C. A. J. Chmura, M. G. Davidson, M. D. Jones, M. D. Lunn, M. F.
Mahon, A. F. Johnson, P. Khunkamchoo, S. L. Roberts, S. S. F.
Wong, Macromolecules 2006, 39, 7250–7257.
[37] J. Balsells, P. J. Carroll, P. J. Walsh, Inorg. Chem. 2001, 40, 5568–
5574.
[38] E. Y. Tshuva, N. Gendeziuk, M. Kol, Tetrahedron Lett. 2001, 42,
6405–6407.
[39] S. Groysman, E. Sergeeva, I. Goldberg, M. Kol, Eur. J. Inorg. Chem.
2005, 2480–2485.
Acknowledgement
We thank Dr. Shmuel Cohen for solution of the X-ray structures. We
also thank HU for funding.
[40] A. Yeori, S. Groysman, I. Goldberg, M. Kol, Inorg. Chem. 2005, 44,
4466–4468.
[1] M. J. Clarke, F. Zhu, D. R. Frasca, Chem. Rev. 1999, 99, 2511–2534.
[2] F. Kratz, M. T. Schꢂtte, Cancer J. 1998, 11, 176–182.
[3] P. Kçpf-Maier, Eur. J. Clin. Pharmacol. 1994, 47, 1–16.
[4] P. Koepf-Maier, H. Kçpf, Chem. Rev. 1987, 87, 1137–1152.
[5] C. V. Christodoulou, A. G. Eliopoulos, L. S. Young, L. Hodgkins,
D. R. Ferry, D. J. Kerr, Brit. J. Cancer 1998, 77, 2088–2097.
[6] F. Caruso, M. Rossi, C. Pettinari, Expert Opin. Ther. Pat. 2001, 11,
969–979.
[7] B. Desoize, Anticancer Res. 2004, 24, 1529–1544.
[8] G. Xu, Y. B. Cui, K. Cui, S. H. Gou, Huaxue Jinzhan 2006, 18, 107–
113.
[9] M. Galanski, V. B. Arion, M. A. Jakupec, B. K. Keppler, Curr.
Pharm. Des. 2003, 9, 2078–2089.
[41] a) Some small changes in chemical shifts were observed following
D2O addition due to the different ratio of solvents. Evaporating the
mixed solution and redissolving in pure [D8]THF gave rise to an
identical spectrum to that of the initial compound; b) the integration
values were calibrated to represent the same number of protons.
[42] K. Matsumoto, Y. Sawada, B. Saito, K. Sakai, T. Katsuki, Angew.
Chem. 2005, 117, 5015–5019; Angew. Chem. Int. Ed. 2005, 44, 4935–
4939.
[43] D. Zhang, Eur. J. Inorg. Chem. 2007, 4839–4845.
[44] T. Kemmitt, N. I. Al-Salim, G. J. Gainsford, Inorg. Chem. 2000, 39,
6067–6071.
[45] A. Kayan, D. Hoebbel, H. Schmidt, J. Appl. Polym. Sci. 2005, 95,
790–796.
[10] I. Ott, R. Gust, Arch. Pharm. Chem. Life Sci. 2007, 340, 117–126.
[11] N. Katsaros, A. Anagnostopoulou, Crit. Rev. Oncol. Hematol. 2002,
42, 297–308.
[46] C. F. Campana, Y. Chen, V. W. Day, W. G. Klemperer, R. A. Sparks,
J. Chem. Soc. Dalton Trans. 1996, 691–702.
[47] G.-J. M. Meppelder, T. P. Spaniol, J. Okuda, J. Organomet. Chem.
2006, 691, 3206–3211.
[48] J. Claffey, M. Hogan, H. Mꢂller-Bunz, C. Pampillon, M. Tacke,
ChemMedChem 2008, 3, 729–731.
[12] A. M. Evangelou, Crit. Rev. Oncol. Hematol. 2002, 42, 249–265.
[13] I. Kostova, Curr. Med. Chem. 2006, 13, 1085–1107.
[14] E. Melendez, Crit. Rev. Oncol. Hematol. 2002, 42, 309.
[15] P. Kçpf-Maier, H. Kçpf, Struct. Bonding 1988, 70, 103–185.
2414
www.chemeurj.org
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 2403 – 2415

TiIV Complexes
FULL PAPER
[49] F. Caruso, C. Pettinari, F. Marchetti, P. Natanti, C. Phillips, J. Tanski,
M. Rossi, Inorg. Chem. 2007, 46, 7553–7560.
[52] SAINT-NT V5.0, Brucker AXS GmbH, Karlsruhe (Germany), 2002.
[53] SHELXTL-NT V6.1, Brucker AXS GmbH, Karlsruhe (Germany),
[50] S. Gendler, A. L. Zelikoff, J. Kopilov, I. Goldberg, M. Kol, J. Am.
Chem. Soc. 2008, 130, 2144–2145.
[51] SMART-NT V5.6, Brucker AXS GmbH, Karlsruhe (Germany),
2002.
2002.
Received: June 30, 2008
Revised: October 6, 2008
Published online: January 20, 2009
Chem. Eur. J. 2009, 15, 2403 – 2415
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
2415