Different ortho and para electronic effects on hydrolysis and cytotoxicity of diamino bis(Phenolato) "Salan" Ti(IV) complexes

Article abstract of DOI:10.1021/ic101693v

Bis(isopropoxo) Ti(IV) complexes of diamino bis(phenolato) "salan" ligands were prepared, their hydrolysis in 1:9 water/THF solutions was investigated, and their cytotoxicity toward colon HT-29 and ovarian OVCAR-1 cells was measured. In particular, electr

Full text of DOI:10.1021/ic101693v

1030 Inorg. Chem. 2011, 50, 1030–1038
DOI: 10.1021/ic101693v
Different ortho and para Electronic Effects on Hydrolysis and Cytotoxicity of
Diamino Bis(Phenolato) “Salan” Ti(IV) Complexes
Dani Peri, Sigalit Meker, Cesar M. Manna, and Edit Y. Tshuva*
Institute of Chemistry, The Hebrew University of Jerusalem, 91904, Jerusalem, Israel
Received August 19, 2010
Bis(isopropoxo) Ti(IV) complexes of diamino bis(phenolato) “salan” ligands were prepared, their hydrolysis in 1:9
water/THF solutions was investigated, and their cytotoxicity toward colon HT-29 and ovarian OVCAR-1 cells was
measured. In particular, electronic effects at positions ortho and para to the binding phenolato unit were analyzed. We
found that para substituents of different electronic features, including Me, Cl, OMe, and NO2, have very little influence
on hydrolysis rate, and all para-substituted ortho-H complexes hydrolyze slowly to give O-bridged clusters with a t1/2 of
1-2 h for isopropoxo release. Consequently, no clear cytotoxicity pattern is observed as well, where the largest
influence of para substituents appears to be of a steric nature. These complexes exhibit IC50 values of 2-18 μM
toward the cells analyzed, with activity which is mostly higher than those of Cp2TiCl2, (bzac)2Ti(OiPr)2 and cisplatin. On
the contrary, major electronic effects are observed for substituents at the ortho position, with an influence that exceeds
even that of steric hindrance. Ortho-chloro or -bromo substituted compounds possess extremely high hydrolytic
stability where no major isopropoxo release as isopropanol occurs for days. In accordance, very high cytotoxicity
toward colon and ovarian cells is observed for ortho-Cl and -Br complexes, with IC50 values of 1-8 μM, where the most
cytotoxic complexes are the ortho-Cl-para-Me and ortho-Br-para-Me derivatives. In this series of ortho-substituted
complexes, the halogen radius is of lesser influence both on hydrolysis and on cytotoxicity, while OMe substituents do
not impose similar effect of hydrolytic stability and cytotoxicity enhancement. Therefore, hydrolytic stability and
cytotoxic activity are clearly intertwined, and thus this family of readily available Ti(IV) salan complexes exhibiting both
features in an enhanced manner is highly attractive for further exploration.
Introduction
the search for other transition metal based drugs revealed
anticancer properties for much of the block d metal groups.4-9
This includes Ti(IV) complexes with various cyclopentadie-
nide or diketonato ligands, which demonstrate cytotoxic
activity toward cisplatin-resistant and -sensitive cells with
reduced side effects (Scheme 1).10-19 However, rapid hydro-
lysis in aqueous environment because of the inherent oxophilic
The interest in metal based anticancer therapeutics is
growing ever since the discovery of the antitumor properties
of cisplatin.1-3 High chelation affinity, attraction to electron
rich molecules, variable coordination number and more,
make it reasonable to assume that metals can mediate
apoptosis through various pathways as can be evident from
the cisplatin anticancer mechanism. However, finite types of
tumors that can be treated with cisplatin and extreme toxic
effects such as neurotoxicity, nephrotoxicity, and others that
usually accompany this chemotherapeutic treatment encou-
rage extensive study of other antitumor tactics, including the
exploration of other metals. Other metal centers can lead to
different antitumor pathways because of different coordina-
tion numbers, electronic characteristics, and so forth. Thus,
(5) Desoize, B. Anticancer Res. 2004, 24, 1529–1544.
(6) Xu, G.; Cui, Y. B.; Cui, K.; Gou, S. H. Prog. Chem. 2006, 18, 107–113.
(7) Ott, I.; Gust, R. Arch. Pharm. Chem. Life Sci. 2007, 340, 117–126.
(8) Bruijnincx, P. C. A.; Sadler, P. J. Curr. Opin. Chem. Biol. 2008, 12,
197–206.
(9) Jakupec, M. A.; Galanski, M.; Arion, V. B.; Hartinger, C. G.;
Keppler, B. K. Dalton Trans. 2008, 183–194.
€
€
€
(10) Kopf-Maier, P.; Kopf, H. Chem. Rev. 1987, 87, 1137–1152.
€
(11) Kopf-Maier, P.; Kopf, H. Struct. Bonding (Berlin) 1988, 70, 103–185.
(12) Keppler, B. K.; Friesen, C.; Moritz, H. G.; Vongerichten, H.; Vogel,
E. Struct. Bonding (Berlin) 1991, 78, 97–127.
(13) Christodoulou, C. V.; Eliopoulos, A. G.; Young, L. S.; Hodgkins, L.;
Ferry, D. R.; Kerr, D. J. Br. J. Cancer 1998, 77, 2088–2097.
(14) Caruso, F.; Rossi, M.; Pettinari, C. Expert Opin. Ther. Pat. 2001, 11,
969–979.
*To whom correspondence should be addressed. Fax þ97226584282.
E-mail: tshuva@chem.ch.huji.ac.il.
(1) Barnes, K. R.; Lippard, S. J. Met. Ions Biol. Syst. 2004, 42, 143–177.
(2) Abu-Surrah, A. S.; Kettunen, M. Curr. Med. Chem. 2006, 13, 1337–
1357.
(3) Cepeda, V.; Fuertes, M. A.; Castilla, J.; Alonso, C.; Quevedo, C.;
Perez, J. M. Anti-Cancer Agents Med. Chem. 2007, 7, 3–18.
(4) Galanski, M.; Arion, V. B.; Jakupec, M. A.; Keppler, B. K. Curr.
Pharm. Des. 2003, 9, 2078–2089.
ꢀ
(15) Melendez, E. Crit. Rev. Oncol. Hematol. 2002, 42, 309–315.
(16) Caruso, F.; Rossi, M. Mini-Rev. Med. Chem. 2004, 4, 49–60.
(17) Kelter, G.; Sweeney, N. J.; Strohfeldt, K.; Fiebig, H.-H.; Tacke, M.
Anti-Cancer Drugs 2005, 16, 1091–1098.
(18) Abeysinghe, P. M.; Harding, M. M. Dalton Trans. 2007, 3474–3482.
(19) Strohfeldt, K.; Tacke, M. Chem. Soc. Rev. 2008, 37, 1174–1187.
r
pubs.acs.org/IC
Published on Web 01/07/2011
2011 American Chemical Society

Article
Inorganic Chemistry, Vol. 50, No. 3, 2011 1031
Scheme 1. Titanocene Dichloride (a) and Budotitane (b)
In the current study we investigated the effect of various
electron donating and withdrawing groups on hydrolysis and
cytotoxicity while comparing the ortho and para effects, with
the assumption that increasing the electron donation ability of
the ligand should provide favored complex features (Chart 1).
Surprising patterns of influence both on hydrolysis and on
cytotoxicity were observed.
Results and Discussion
nature of Ti(IV) ions combined with electron poor configura-
tion is the Achilles’ heel of these compounds, which in turn led
to failure in clinical trials.12,13,20-23 In the presence of water,
the labile groups (Cl, OEt) of titanocene dichloride (Cp2TiCl2,
Scheme 1, a) and budotitane ((bzac)2Ti(OEt)2, Scheme 1, b)
are hydrolyzed within minutes, followed by the more inert
ligands (Cp, diketonato) within hours, leading to unidentified
aggregates which hamper further mechanistic investigations.
However, extensive studies of these compounds show tunable
features of the anticancer properties.24 Thus, a great challenge
remains in the design and synthesis of better-suited Ti(IV)
complexes that would demonstrate enhanced hydrolytic sta-
bility and improved antitumor properties.
Salans are well-known diamine bis(phenolato) compounds
which have been used as chelating ligands for a wide variety of
transition metals for various applications. We recently intro-
duced the Ti(IV) salan complexes as a new family of highly
cytotoxic compounds,24-27 with cytotoxicity greater than that
of Cp2TiCl2, (bzac)2Ti(OiPr)2, and cisplatin toward colon and
ovarian cells. We have demonstrated that their particularly
slow hydrolysis enables detailed studies in water-enriched
environments, as the labile isopropoxo ligands hydrolyze
within hours in 1:9 water/THF solutions, and the resulting
salan-bound cluster does not further hydrolyze for days. We
have detected a strong correlation between hydrolysis pathway
and cytotoxicity, where formation of polynuclear O-bridged
hydrolysis products were obtained for all cytotoxic complexes,
and inactive complexes mostly released the free bis(phenol)
ligand. Moreover, we have established that strong ligand
binding enables its interaction with the cellular target as was
particularly evident by a stereochemical study.28 Furthermore,
study of different derivatives of different steric and electronic
properties by us,24-27 followed by others,29 have demonstrated
that large steric bulk is of negative influence on cytotoxic
activity, both when located near the metal center and when
located on a peripheral area of the complex. This supports the
notion that planarity is of importance for DNA intercalation;
if indeed DNA is the cellular target.12,28
In the current study we attempted to isolate electronic
effects as much as possible and study their influence on the
complex activity in particular at the ortho and at the para
positions, with the notion that increasing the electron dona-
tion ability of the ligand should strengthen its metal binding
and thus increase the complex hydrolytic stability, and
consequently, its cytotoxicity. Synthesis of most ligand pre-
cursors L1,2,4-8H2 is achieved via a single step procedure
according to known procedures,30 from the commercially
available diamine, substituted phenol, and formaldehyde.
L3Ti(OiPr)2 is synthesized differently because of the electron
withdrawal nature of the substituent, by a nucleophilic
substitution on 2-chloromethyl-4-nitrophenol according to
a published procedure.31 Synthesis of Ti(IV) complexes was
performed in analogy to known compounds under an inert
atmosphere by mixing the ligand L1-8H2 with 1 equiv of
Ti(OiPr)4 in tetrahydrofuran (THF) at room temperature
(RT) and stirring for about 2 h.32-34 Removing the solvent
under reduced pressure produced the product in quantitative
yield, which may then be further recrystallized from diethy-
lether to give yellow crystals or a yellow crystalline powder
(Scheme 2).
Complexes L1-8Ti(OiPr)2 (Chart 1) all share similar struc-
tural features to known complexes of this type,25,32,33,35,36 as
evident mainly by NMR featuring the two AX systems of the
methylene protons. They thus exhibit symmetry of C2 be-
cause of trans binding of the phenolato donors and a cis
configuration of the isopropoxo groups. A representative
complex L5Ti(OiPr)2 was also analyzed by single crystal
X-ray crystallography (Figure 1). A list of selected bond
lengths and angles for this complex is given in Table 1. When
comparing the structure of L5Ti(OiPr)2 to that of its analo-
gues with only alkyl substituents on the aromatic rings,
namely, with methyl groups meta and para to the phenolato
oxygen atoms,26 it is evident that the chloro substituents do
not have a major effect on the coordination sphere of the
Ti(IV) ion. For instance, the Ti-O1 and Ti-O2 distances
˚
vary between 1.91 and 1.93 A, relative to a parallel value of
˚
1.90 A in the alkylated complex, and the O1-Ti-O2 angle of
163.0° is only slightly smaller than the 167.1° angle observed
(20) Toney, J. H.; Marks, T. J. J. Am. Chem. Soc. 1985, 107, 947–953.
(21) Schilling, T.; Keppler, K. B.; Heim, M. E.; Niebch, G.; Dietzfelbinger,
H.; Rastetter, J.; Hanauske, A. R. Invest. New Drugs 1995, 13, 327–332.
(22) Korfel, A.; Scheulen, M. E.; Schmoll, H. J.; Grundel, O.; Harstrick,
A.; Knoche, M.; Fels, L. M.; Skorzec, M.; Bach, F.; Baumgart, J.; Sass, G.;
Seeber, S.; Thiel, E.; Berdel, W. E. Clin. Cancer Res. 1998, 4, 2701–2708.
(23) Caruso, F.; Massa, L.; Gindulyte, A.; Pettinari, C.; Marchetti, F.;
Pettinari, R.; Ricciutelli, M.; Costamagna, J.; Canales, J. C.; Tanski, J.;
Rossi, M. Eur. J. Inorg. Chem. 2003, 3221–3232.
(24) Tshuva, E. Y.; Ashenhurst, J. A. Eur. J. Inorg. Chem. 2009, 2203–
2218.
(25) Shavit, M.; Peri, D.; Manna, C. M.; Alexander, J. S.; Tshuva, E. Y.
J. Am. Chem. Soc. 2007, 129, 12098–12099.
(26) Peri, D.; Meker, S.; Shavit, M.; Tshuva, E. Y. Chem.;Eur. J. 2009,
15, 2403–2415.
(27) Tshuva, E. Y.; Peri, D. Coord. Chem. Rev. 2009, 253, 2098–2115.
(28) Manna, C. M.; Tshuva, E. Y. Dalton Trans. 2010, 39, 1182–1184.
(29) Immel, T. A.; Groth, U.; Huhn, T. Chem.;Eur. J. 2010, 16, 2775–
2789.
in the alkylated analogue.
(30) Tshuva, E. Y.; Gendeziuk, N.; Kol, M. Tetrahedron Lett. 2001, 42,
6405–6407.
(31) Whiteoak, C. J.; Britovsek, G. J. P.; Gibson, V. C.; White, A. J. P.
Dalton Trans. 2009, 2337–2344.
(32) Balsells, J.; Carroll, P. J.; Walsh, P. J. Inorg. Chem. 2001, 40, 5568–
5574.
(33) Chmura, A. J.; Davidson, M. G.; Jones, M. D.; Lunn, M. D.; Mahon,
M. F.; Johnson, A. F.; Khunkamchoo, P.; Roberts, S. L.; Wong, S. S. F.
Macromolecules 2006, 39, 7250–7257.
(34) Gendler, S.; Segal, S.; Goldberg, I.; Goldschmidt, Z.; Kol, M. Inorg.
Chem. 2006, 45, 4783–4790.
(35) Groysman, S.; Sergeeva, E.; Goldberg, I.; Kol, M. Eur. J. Inorg.
Chem. 2005, 2005, 2480–2485.
(36) Yeori, A.; Groysman, S.; Goldberg, I.; Kol, M. Inorg. Chem. 2005,
44, 4466–4468.

1032 Inorganic Chemistry, Vol. 50, No. 3, 2011
Chart 1. Salan Complexes Studied
Peri et al.
5
ꢀ
˚
Table 1. Selected Bond Lengths (A) and Angles (deg) for L Ti(OiPr)2
atoms
value
atoms
value
Lengths
O(1)-Ti
1.911(3)
1.932(3)
1.791(3)
1.806(3)
N(1)-Ti
N(2)-Ti
2.337(3)
2.345(3)
O(2)-Ti
O(3)-Ti
O(4)-Ti
Figure 1. ORTEP drawingofL5Ti(OiPr)2 in50% probability ellipsoids.
H atoms were omitted for clarity.
Angles
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)
106.08(13)
92.73(12)
97.67(13)
96.62(12)
93.29(12)
163.05(11)
76.08(11)
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)
164.68(12)
88.43(12)
80.19(11)
87.26(11)
89.85(12)
163.65(12)
85.01(12)
80.93(11)
Scheme 2. Synthesis of Diamino Bis(phenolato) Salan Complexes
[HONNOH], and a new product of hydrolysis, were ana-
lyzed over time to provide information on the hydrolysis
pathway, including reaction products and hydrolysis rates. A
representative plot for L1Ti(OiPr)2 is shown in Figure 2, and
a summary of t1/2 values of hydrolysis of the isopro-
poxo groups for L1-8Ti(OiPr)2 is provided in Table 2, as
calculated from the plots of isopropoxo release presented in
Figure 3.
In an attempt to exclude steric effects as much as possible,
in the series of L1-4Ti(OiPr)2 (Chart 1) the complexes were
substituted only at the para position with substituents of
markedly different electronic features. It is noteworthy that
the steric effects at this position were also found in our
previous studies to affect cytotoxicity.26 The half-life values
of isopropoxo hydrolysis for complexes L1-4Ti(OiPr)2 is in
the same order of magnitude and varied between 1 and 2 h
(Table 2, Figure 2,3), which corresponds to stability that
generally exceeds that of titanocene dichloride and
Hydrolysis Studies. As analogous complexes are gen-
erally stable for at least several hours in water enriched
environments,26 we were interested to establish the rela-
tive water resistance of this series of complexes as well.
Therefore, we employed the NMR technique to monitor the
hydrolysis reaction that occurs following D2O addition to a
d8-THF solution of the L1-8Ti(OiPr)2 (Chart 1) to give a 1:9
D2O/d8-THF solution. The spectrum was measured every
5-10 min for up to 60 h, and the integration of selected
signals was measured to gain insight on the complexes’
decomposition process. In particular, the half-life of the
isopropoxo ligands hydrolysis was measured by monitoring
the integration of the septet signal of the CH group and/or
the doublet of the methyl groups of the bound isopropoxo
ligands. Other signals representing bound phenolato ligands
in the original complex [Ti-ONNO], free bis(phenol)

Article
Inorganic Chemistry, Vol. 50, No. 3, 2011 1033
overtime (Figure 2) confirms that this cluster is the major
hydrolysis product obtained simultaneously to isopropoxo
release, based on very little formation of free salan ligand,
<20%.37 It thus seems that the hydrolytic stability of these
complexes is derived mostly from the inherent nature of the
phenolato moieties, and is more dramatically affected by
steric groups ortho to the metal binding site that affect
formation of a cluster by imposing a kinetic barrier,26 rather
than by electronic effects on the periphery of the aromatic
rings which are negligible. Therefore, complexes that are
unsubstituted at the ortho positions such as L1-4Ti(OiPr)2
all give similar hydrolysis rates.
In a previous study, we reported that LaTi(OiPr)2
(Chart 1) is remarkably stable with a half-life for isopropoxo
hydrolysis of 31 h, with slow formation of a cluster because
of the large ortho methyl substituent imposing a kinetic
barrier (Figure 3).26 We were thus interested to examine a
different series of complexes, with substituents of different
electronic features at the ortho position as well.
Figure 2. Representative plot of the integration of selected signals in the
1H NMR spectrum of L1Ti(OiPr)2 versus time following addition of D2O
to a d8THF solution of the complex at RT. All signals except for the “new
species” were calibrated to represent a single proton. It is concluded that
the new species is of reduced symmetry, and the selected signal represents
reduced number of protons (vide infra).
Hydrolysis studies conducted for L5-7Ti(OiPr)2 (Chart 1)
indicated remarkable stability, substantially higher than that
of related complexes that do not include ortho Cl or Br
substituents,25 despite the very small effect of the Cl substit-
uents on the structure of L5Ti(OiPr)2 as evident by its crystal
structure (Figure 1, Table 1). Monitoring the integration of
1
the signals in the H NMR of complexes L5-7Ti(OiPr)2
following D2O addition over a period of 60 h showed
extremely slow change in integration of both the bound
phenolato and the presumably more labile isopropoxo
ligands, (Figure 3, Table 2), and new signals of a hydrolysis
product could barely be detected (Figure 4). It is thus
obvious that these complexes do not undergo significant
hydrolysis for days under these conditions.
L6Ti(OiPr)2 (Chart 1) featuring only ortho-Cl groups
demonstrates similar hydrolytic stability to that of L5Ti-
(OiPr)2, featuring both ortho- and para-Cl substituents
(Table 2, Figure 3). This supports the conclusion that
electronic effects at the para position are of lesser sig-
nificance as observed for L1-4Ti(OiPr)2. Adding L7Ti-
Figure 3. Plots of the integration of the Ti-OiPr signal in the 1H NMR
spectra of the complexes studied versus time following addition of D2O to
a d8-THF solution of the complex at RT.
(OiPr)2 and LaTi(OiPr)2 (Chart 1) to the discussion
26
points to this effect being mainly, but not solely, electronic,
as L7Ti(OiPr)2 demonstrates even higher hydrolytic stability
than that of L5,6Ti(OiPr)2 presumably because of the larger
volume of the Br substituents, and the stability of all three
L5-7Ti(OiPr)2 is even higher than that of LaTi(OiPr)2
featuring solely Me substituents (Table 2, Figure 3). Inter-
estingly, the markedly enhanced hydrolytic stability is not
observed for the methoxy substituted complex L8Ti(OiPr)2
with a half-life of isopropoxo hydrolysis of 3.4 h (Table 2,
Figure 3) to give a cluster as the main product (>70%),
despite its greater steric bulk. This supports the conclusion
that this effect of hydrolytic stability enhancement in mainly
electronic.
Table 2. t1/2 (h) Values of Isopropoxo Release of L1-8Ti(OiPr)2 at 1:9 D2O/d8-
THF Solution at RT
complex
substituents
t1/2
L1Ti(OiPr)2
L2Ti(OiPr)2
L3Ti(OiPr)2
L4Ti(OiPr)2
L5Ti(OiPr)2
L6Ti(OiPr)2
L7Ti(OiPr)2
L8Ti(OiPr)2
o-H-p-Me
o-H-p-Cl
o-H-p-NO2
o-H-p-OMe
o-Cl-p-Cl
o-Cl-p-Me
o-Br-p-Me
o-OMe-p-Me
2.1 ( 0.8
2.1 ( 0.6
1.3 ( 0.7
2.0 ( 0.2
110 ( 10a
130 ( 30a
290 ( 30a
2.9 ( 0.4
a Estimated based on extrapolation.
Because of the observed high water stability of L5-7
-
budotitane.12,15,20 Surprisingly, it appears that the electronic
properties of the substituents in the para position, including
such that are able to impose resonative effects, play little role
Ti(OiPr)2, these complexes were reacted with varying equi-
valents of water: 50, 100, and 1000, over a 3 days period, in
an attempt to structurally characterize hydrolysis products.
1
in this high stability. In addition, the H NMR spectra of
L1-4Ti(OiPr)2 taken several hours following D2O addition
includes many new signals of a new species of reduced
symmetry, presumed to be a polynuclear o-bridged com-
pound with bound salan ligands as observed for related
complexes (Figure 4).26 Moreover, the integration change
(37) As the polynuclear hydrolysis products are of low symmetry, the
multiple signals of low integration in their 1H NMR spectra make it difficult
to assign complete proton identity. Therefore, the percentage of cluster
formed is estimated based on the percentage of free bis(phenol) ligand
released.

1034 Inorganic Chemistry, Vol. 50, No. 3, 2011
Peri et al.
Figure 4. 1H NMR spectra of L1Ti(OiPr)2 (left) and L7Ti(OiPr)2 (right) at RT in d8-THF immediately (bottom) and 20 h (top) following addition of D2O.
6
ꢀ
˚
Table 3. Selected Bond Lengths (A) and Angles (deg) for Ti3L 3(μ-O)3, the
Hydrolysis Product of L6Ti(OiPr)2
atoms
value
atoms
value
Lengths
O(1)-Ti(1)
O(3)-Ti(1)
O(4)-Ti(1)
O(5)-Ti(1)
O(1)-Ti(2)
O(2)-Ti(2)
O(2)-Ti(3)
O(3)-Ti(3)
O(6)-Ti(2)
O(7)-Ti(2)
O(8)-Ti(3)
O(9)-Ti(3)
1.831(2)
1.793(2)
1.943(2)
1.934(2)
1.805(2)
1.872(2)
1.801(2)
1.913(2)
1.954(2)
1.871(2)
1.938(2)
1.875(2)
N(1)-Ti(1)
N(2)-Ti(1)
N(3)-Ti(2)
N(4)-Ti(2)
N(5)-Ti(3)
N(6)-Ti(3)
2.337(3)
2.368(3)
2.258(3)
2.361(3)
2.281(3)
2.354(3)
Figure 5. ORTEP drawing of Ti3L63(μ-O)3, the hydrolysis product of
L6Ti(OiPr)2, in 50% probability ellipsoids. H atoms and disordered
solvent were omitted for clarity.
Angles
O(3)-Ti(1)-O(1)
O(1)-Ti(2)-O(2)
O(2)-Ti(3)-O(3)
O(1)-Ti(2)-O(7)
O(7)-Ti(2)-O(2)
O(1)-Ti(2)-O(6)
O(7)-Ti(2)-O(6)
O(2)-Ti(2)-O(6)
Ti(1)- O(1)- Ti(2)
Ti(3)- O(2)- Ti(2)
Ti(1)- O(3)- Ti(3)
101.12(9)
92.94(9)
95.61(9)
109.47(10)
97.40(10)
91.78(9)
90.25(10)
169.12(10)
141.47(12)
146.66(12)
135.55(11)
O(1)-Ti(2)-N(3)
O(7)-Ti(2)-N(3)
O(2)-Ti(2)-N(3)
O(6)-Ti(2)-N(3)
O(1)-Ti(2)-N(4)
O(7)-Ti(2)-N(4)
O(2)-Ti(2)-N(4)
O(6)-Ti(2)-N(4)
N(3)-Ti(2)-N(4)
93.77(9)
155.11(10)
89.96(9)
79.95(9)
170.52(10)
79.84(10)
87.48(10)
86.26(10)
76.76(10)
Single crystals of the isolated powder obtained from the
reaction of L5Ti(OiPr)2 with 100 water equivalents were
analyzed by X-ray crystallography, and the structure fea-
tures the starting complex L5Ti(OiPr)2 (Figure 1). This is
1
consistent with the H NMR spectrum of the product
indicating that L5Ti(OiPr)2 is the main compound present
in solution, although traces of a polynuclear product could
also be detected (Supporting Information, Figure S8). This
supports our conclusion regarding the high hydrolytic
stability of this compound.29 Interestingly, however, the
reaction of L6Ti(OiPr)2 with 50 water equivalents yielded
single crystals of the extremely minor polynuclear product
reported formed from an ortho-H-para-Me-meta-Me deri-
vative,26 which supports the notion that similar hydrolysis
pathways occur for much of the complexes of this family
with different steric and electronic demands. It is thus clear
that the especially stable ortho-halogenated compounds are
also able to form the O-bridged hydrolysis products over
long periods and/or under conditions of high water con-
centrations. A comparison of the 1H NMR spectra of the
hydrolysis products obtained for L5-7Ti(OiPr)2 reacted
with 1000 water equivalents for 3 days is presented in Figure
6, which demonstrates the relative portion of polynuclear
compounds obtained for these complexes and the particu-
larly enhanced stability of L7Ti(OiPr)2. Additional 1H
NMR spectra of the products obtained following addition
of 50 and 100 water equivalents are given in Supporting
Information, Figure S7-S12.
1
as evident by H NMR (Supporting Information, Figure
S9). The crystals contain disordered diethylether and water
molecules. Nevertheless, the structure of the main molecule
(Figure 5, Table 3) features a trinuclear Ti(IV) complex
bridged by three oxo units, with a phenolato ligand bound
to each Ti(IV) center. One out of the three Ti(IV) cores
(Ti(1)) retains the original trans configuration, while the
other two (Ti(2), Ti(3)) feature cis binding of the phenolato
units, giving altogether a symmetry of C1. This low sym-
metry explains the multiple signals observed in the 1H NMR
for new hydrolysis products (Figure 4). The nearest Cl Cl
3 3 3
˚
distance for two different ligands is 3.7 A. This structure is
generally highly similar to another cluster we previously

Article
Inorganic Chemistry, Vol. 50, No. 3, 2011 1035
Figure 6. 1H NMR spectra of L5Ti(OiPr)2 (bottom), L6Ti(OiPr)2 (middle), and L7Ti(OiPr)2 (top) 3 days following addition of 1000 water equivalents; the
integration of remaining bound isopropoxo relative to a new doublet of a polynuclear hydrolysis product is presented.
Table 4. Relative IC50 (μM) and Maximal Cell Growth Inhibition (%) of
L
1-8Ti(OiPr)2 Towards HT-29 and OVCAR-1 Cancer Cell Lines
complex
HT-29
OVCAR-1
L1Ti(OiPr)2
L2Ti(OiPr)2
L3Ti(OiPr)2
L4Ti(OiPr)2
L5Ti(OiPr)2
L6Ti(OiPr)2
L7Ti(OiPr)2
L8Ti(OiPr)2
Cp2TiCl2
6.2 ( 0.9 (90%)
2.9 ( 0.5 (80%)
2.6 ( 0.6 (83%)
16.6 ( 3.8 (82%)
2.4 ( 0.4 (51%)
7.2 ( 2.0 (76%)
2.0 ( 0.5 (70%)
8.1 ( 1.6 (98%)
609 ( 4 (90%)
15.2 ( 0.3 (92%)
5.2 ( 1.1 (79%)
11.5 ( 2.7 (82%)
3.3 ( 0.8 (70%)
17.7 ( 3.6 (76%)
2.7 ( 0.5 (52%)
4.5 ( 0.9 (83%)
1.5 ( 0.3 (69%)
6.3 ( 0.9 (98%)
701 ( 4 (91%)
(bzac)2Ti(OiPr)2
14.9 ( 0.4 (91%)
We may thus conclude that the ortho substituents are of
the largest influence on hydrolytic stability, which may
impose not only steric26 but also significant electronic
effects. Consequently, as high water resistance is observed
for the complexes presented herein, and yet, their ability
to form clusters over long periods was established, both
features proved essential for high antitumor activity in
our previous studies,26 we were eager to continue to
cytotoxicity measurements.
Cytotoxicity Measurements. The cytotoxicity measure-
ments were carried out with two types of cell lines: ovarian
OVCAR-1 and colon HT-29, and analysis was achieved
by the MTT (methylthiazolyldiphenyl-tetrazolium bromide)
assay following 3d incubation of the cells with the complex
analyzed. A summary of relative IC50 values and maximal
cell growth inhibition are presented in Table 4, and repre-
sentative plots of cell viability versus concentration toward
both cell types are given in Figure 7.
Figure 7. Dependence of HT-29 (top) and OVCAR-1 (bottom) cell
viability after 3 d incubation period based on the MTT assay on
administered concentration of L2,3,7Ti(OiPr)2 presented in logarithmic
scale.
In general, we observe that the correlation between hy-
drolytic behavior and stability and the cytotoxic acti-
vity that we reported on previously is retained.26 Com-
plexes L1-4Ti(OiPr)2 which are substituted solely at the

1036 Inorganic Chemistry, Vol. 50, No. 3, 2011
Peri et al.
para positions of the aromatic moieties (Chart 1), exhibit
high cytotoxic activities with IC50 values between 3 and
18 μM (Table 4), activity which is also mostly higher than
that of Cp2TiCl2, (bzac)2Ti(OiPr)2, and cisplatin toward
the cells analyzed.25 This stands in agreement with the
hydrolytic behavior of L1-4Ti(OiPr)2, involving relatively
slow hydrolysis to give polynuclear products, features
that were shown in our previous studies to be essential for
cytotoxic activity. In addition, as the hydrolysis is not
largely influenced by the substituent electronic features,
no clear dependence pattern of the cytotoxicity on the
electronic donation ability of the para substituents can be
detected, where in fact, the most active complex in this
series of L1-4Ti(OiPr)2 is L3Ti(OiPr)2 with a IC50 values
of 2.6 ( 0.6 and 3.3 ( 0.8 μM for the colon and ovarian
cells, respectively (Table 4), which bears the most electron
withdrawing groups of all, NO2 (Chart 1). One possible
reason to the highest activity of L3Ti(OiPr)2 may relate to
the high planarity of the NO2 substituents and their “flat”
nature, which might support DNA intercalation taking
part in the activity mechanism, and is in agreement with
our previous observation regarding negative effect of
large para substituents such as t-Bu on cytotoxicity.26
On the same note, L1,4Ti(OiPr)2 exhibit IC50 values of
6.2 ( 0.9 (colon) and 5.2 ( 1.1 (ovarian), and 16.6 ( 3.8
(colon) and 17.7 ( 3.6 (ovarian) μM, respectively
(Table 4). This activity difference may again be attributed
to steric hindrance and interruption to planarity, rather
than to different electronic effects.
The complexes halogenated at the ortho positions
demonstrate exceptionally high cytotoxicity particularly
at low concentrations with the IC50 values varying be-
tween 2 and 7 μM for colon cells and 1-5 μM for ovarian
cells, which may relate to their especially high hydrolytic
stability.29,38 For this series, the steric effects seem to be
relatively minor, as L5Ti(OiPr)2 is more active than LaTi-
(OiPr)2 (Chart 1), and even higher activity is observed for
L6,7Ti(OiPr)2 when taking maximal inhibition into con-
sideration (Table 4). These observations may suggest that
the electronic properties induced at the ortho position
have larger effect on the cytotoxicity than the steric
hindrance, perhaps because of their effect on hydrolytic
stability. Furthermore, placing a substituent with higher
steric hindrance such as a methoxy group at the ortho
position (L8Ti(OiPr)2, Chart 1) produces a species with
similar activity to that of LaTi(OiPr)2.
and at the para positions, which are largely available through a
highly convenient synthetic pathway, making this family of
complexes especially attractive for study and development.
High activity is observed for complexes of this family on
OVCAR-1 and HT-29 cells, higher than that of Cp2TiCl2,
(bzac)2Ti(OiPr)2, and cisplatin, which is accompanied by
exceptionally high hydrolytic stability, higher than that of
known Ti(IV) cytotoxic complexes. In addition, we con-
tinue to observe the relation between these parameters:
highly hydrolytically stable complexes lead to high cyto-
toxicity. We observe that the active salan complexes very
slowly give polynuclear clusters with bound phenolato
ligand upon hydrolysis, involving release of the labile
isopropoxo ligands as isopropanol. This is not always the
case for inactive complexes, which largely release free bis-
(phenol) ligands because of steric hindrance or inherent
hydrolytic instability.26 We overall continue to observe a
big influence of the particular ligand on the complex perfor-
mance, which certainly encourages further ligand design.
We found that contrary to our assumption, varying
electronic features of para substituents have very little influ-
ence on hydrolytic stability, identifying the diamine bis-
(phenolato) core as the main responsible parameter for the
inherent water resistance of these complexes. The cytotoxi-
city, on the other hand, seems to depend much more on steric
effects and the planarity of the substituents, which may be
attributed to the involvement of DNA intercalation in the
activity mechanism,12 which is also in agreement with our
recent report identifying the biological target as chiral.28
Nevertheless, influence of these parameters on cell penetra-
tion cannot be ruled out.
We observed strong positive electronic influence of ortho-
Cl and -Br substituents on hydrolytic stability, which is not
observed for para substitution, the reason for which is not
exactly clear. In particular, the para-methylated-ortho-haloge-
nated complexes we prepared are extremely stable with no
substantial release of the most labile isopropoxo ligands for
days in 1:9 water/THF solutions. Moreover, these complexes
demonstrate higher cytotoxicity than their analogues, with
some effect of the halogen radius on hydrolysis and cytotoxi-
city. In addition, the eventual formation of polynuclear
hydrolysis products, established based on X-ray crystallogra-
phy for one compound and supported by NMR analyses
of all, is in agreement with our previous studies suggesting
that the steric ability to form such clusters is essential for
cytotoxicity. We thus suspect one of two options: (a) the
steric requirement to form the cluster is similar to that of the
interaction with the biological target; (b) the original bis-
(isopropoxo) complex is required for cell penetration, and
the cluster itself or an intermediate in its formation might be
active inside the cell. Indeed, all highly cytotoxic complexes of
the series studied herein form such clusters over long periods in
water environment. Nevertheless, as the clusters themselves
are inactive,26 either in terms of cell penetration or of cellular
interactions, it is concluded that decreasing the rate of hydro-
lysis to give them is of importance for enhancing cytotoxic
activity of diamino bis(phenolato) “salan” Ti(IV) complexes.
Therefore, the para-Me-ortho-Br complex L7Ti(OiPr)2 exhibits
the highest cytotoxic activity of all, accompanied by incredibly
high water stability, and represents the most promising anti-
tumor Ti(IV) complex of this family identified thus far.
In conclusion, electronic variation enabled as to isolate
readily available salan Ti(IV) complexes of especially high
Conclusions
In this study we were interested to evaluate whether
controlled substitutions with electron withdrawing and do-
nating groups at the aromatic moieties of the tetradentate
ligands at the ortho and at the para positions to the metal
binding site can lead to additional insight on the hydrolytic
behavior and the biological activity of this promising new
family of cytotoxic Ti(IV) compounds that we recently intro-
duced. We worked under the assumption that increasing
electron donation ability of the ligands should also increase
hydrolytic stability, and therefore, might affect cytotoxicity as
well. We thus analyzed several complexes with varying sub-
stituents of different electronic features positioned at the ortho
(38) Immel, T. A.; Debiak, M.; Groth, U.; Burkle, A.; Huhn, T.
ChemMedChem 2009, 4, 738–741.

Article
Inorganic Chemistry, Vol. 50, No. 3, 2011 1037
hydrolytic stability and cytotoxicity. Additional mechanistic
studies relating to the hydrolysis pathway and its relation to
the cytotoxicity mechanism are underway.
solvent alone. In some cases, a decrease in activity is observed
at high concentration, which may be due to enhanced aggrega-
tion to give inactive polynuclear products.
L6H2. A solution of 2-chloro-4-methylphenol (0.59 mL,
5.0 mmol), N,N0-dimethylethylenediamine (0.27 mL, 2.5 mmol)
and paraformaldehyde (0.30 g, 12 mmol) in methanol (5 mL)
was stirred and refluxed for 24 h. The mixture was cooled and
the colorless precipitate was filtered and washed with cooled
methanol (0.85 g, 86%). 1H NMR (CDCl3) δ 7.07 (2H, d, J =
2.1 Hz, Ar), 6.66 (2H, d, J = 1.9 Hz, Ar), 3.67 (4H, s, CH2), 2.69
(4H, s, CH2), 2.30 (6H, s, CH3), 2.21 (6H, s, CH3); 13C NMR
(CDCl3) δ: 151.1, 129.6, 129.1, 127.6, 122.3, 120.4, 61.6, 54.3,
41.9, 20.3; Anal. Calcd for C20H26Cl2N2O2: C, 60.46; H, 6.60; N,
7.05. Found: C, 60.49; H, 6.74; N, 7.04.
Experimental Section
Ligands and their bis(isopropoxo) Ti(IV) complexes were
synthesized according to published procedures.30-34 Data on
H2L1-4 31 H2L5,34 H2L8,39 L3Ti(OiPr)2,33 L5Ti(OPr)2,34 and
,
L8Ti(OiPr)239 can be found elsewhere. Paraformaldehyde, N,
N0-dimethylethylenediamine and all substituted phenol com-
pounds were purchased from Aldrich Chemical Co. Inc. or
€
Fluka Riedel-deHaen. Titanium tetra(isopropoxide) (97%)
was purchased from Aldrich Chemical Co., Inc. All solvents
weredistilled fromK or K/benzophenone under nitrogen. All
experiments requiring dry atmosphere were performed in a
M. Braun drybox or under nitrogen atmosphere using
Schlenk line techniques. NMR data were recorded using
AMX-400 MHz or AMX-500 MHz Bruker spectrometer.
X-ray diffraction data were obtained with a Bruker SMART
APEX CCD diffractometer, running the SMART software
package. After collection, the raw data frames were inte-
grated by the SAINT software package. The structures were
solved and refined using the SHELXTL software package.
Elemental analyses were performed in the microanalytical
laboratory in our institute. Kinetic studies by NMR to
monitor hydrolysis of isopropoxo groups to give polynuclear
products, for the establishment of relative water resistance,
were performed using about 6 mM of the complex solution in
d8-THF and adding D2O to give a final solution of 1:9 D2O/
d8-THF, withaddedD2Obeing>1000equivalentsrelativeto
Ti(IV). The average t1/2 value based on a pseudo first order fit
(Supporting Information, Figure S2) of three replicates is
reported for each compound. The results were verified by
including p-dinitro benzene as an internal standard. The sum
of integration of iPrOH and Ti-OiPr in the first measurement
following D2O addition was assigned as integration 1. At-
tempts to crystallize the hydrolysis products were performed
using 2-20mMofthe complex as a THF solution andadding
50, 100, or 1000 water equivalents.
L7H2. was synthesized similarly from 2-bromo-4-methylphenol
(0.24 mL, 2.0 mmol), N,N0-dimethylethylenediamine (0.11 mL, 1.0
mmol), and paraformaldehyde (0.09 g, 3.0 mmol) (0.31 g, 63%).
1H NMR (CDCl3) δ 7.24 (2H, m, Ar), 6.71 (2H, m, Ar), 3.66 (4H,
s, CH2), 2.70 (4H, s, CH2), 2.30 (6H, s, CH3), 2.22 (6H, s, CH3);13
C
NMR (CDCl3) δ152.0, 132.5, 129.6, 128.3, 122.2, 109.8, 61.6, 54.3,
41.8, 20.2; Anal. Calcd for C20H26Br2N2O2: C, 49.40; H, 5.39; N,
5.76. Found: C, 49.64; H, 5.46; N, 5.52.
L1Ti(OiPr)2. Ti(OiPr)4 (0.050 g, 0.18 mmol) was dissolved in
5 mL of dry THF under inert atmosphere. L1H2 (0.065 g,
0.18 mmol) was dissolved in 10 mL of dry THF under inert
atmosphere. The two solutions were combined and allowed to mix
under ambient conditions at room temperature for 2 h. The solvent
was removed with reduced pressure to give the product in a
quantitative yield, which may then be recrystallized from diethy-
lether (0.028 g, 32%). 1H NMR (CDCl3) δ 6.96 (2H, dd, J = 8.1,
2.2 Hz, Ar), 6.74 (2H, d, J = 1.9 Hz, Ar), 6.57 (2H, d, J = 8.1 Hz,
Ar), 5.02 (2H, sept, J = 6.1 Hz, OCH(CH3)2), 4.62 (2H, d, J =
13.4 Hz, CH2), 3.06(2H, d, J= 13.5 Hz, CH2), 2.99 (2H, d, J=9.3
Hz, CH2), 2.44 (6H, s, CH3), 2.24 (6H, s, CH3), 1.76(2H, d, J=9.4
Hz, CH2), 1.24 (12H, d, J = 6.1 Hz, OCH(CH3)2); 13C NMR
(CDCl3) δ 159.8, 129.9, 129.5, 126.4, 124.3, 117.2, 77.6, 64.5, 51.8,
47.2, 26.0, 25.7, 20.5; Anal. Calcd for C26H40N2O4Ti: C, 63.41; H,
8.19; N, 5.69. Found: C, 63.41; H, 8.05; N, 5.49.
L2Ti(OiPr)2. was synthesized similarly in quantitative yield
from Ti(OiPr)4 (0.050 g, 0.18 mmol) and L2H2 (0.069 g, 0.18
mmol), and may be recrystallyzed from diethylether (0.039 g,
Cytotoxicity was measured on HT-29 colon and OVCAR-1
ovarian cells obtained from ATCC Inc. using the methylthia-
zolyldiphenyl-tetrazolium bromide (MTT) assay. In more
detail: cells (1.2ꢀ106) in medium (contains: 1% penicillin/
1
41%). H NMR (CDCl3) δ 7.10 (2H, dd, J = 8.6, 2.6 Hz, Ar),
6.92 (2H, d, J=2.6Hz, Ar), 6.59(2H, d,J= 8.6 Hz, Ar), 4.96 (2H,
sept, J = 6.2 Hz, OCH(CH3)2), 4.57 (2H, d, J = 13.8 Hz, CH2),
3.09 (2H, d, J= 13.6 Hz, CH2), 2.96 (2H, d, J=9.4Hz, CH2), 2.43
(6H, s, CH3), 1.84 (2H, d, J = 9.4 Hz, CH2), 1.23 (6H, d, J = 6.1
Hz, OCH(CH3)2), 1.22 (6H, d, J = 6.1 Hz, OCH(CH3)2); 13C
NMR (CDCl3) δ 160.7, 129.0, 125.9, 121.7, 118.8, 78.4, 77.2, 63.9,
51.8, 47.2, 25.9, 25.6; Anal. Calcd for C24H34Cl2N2O4Ti: C, 54.05;
H, 6.43; N, 5.25. Found: C, 54.07; H, 6.57; N, 5.08.
streptomycin antibiotics; 1% L-glutamine; 10% fetal bovine
serum (FBS), all purchased from Biological Industries Inc.,
and 88% medium RPMI-1640, purchased from Sigma Inc.)
were seeded into a 96-well plate and allowed to attach for 24 h.
The cells were consequently treated with the reagent tested at
10 different concentrations. Solution of reagent was prepared
by dissolving the reagent in 10 μL of the suitable solvent (THF
or H2O) and diluting with 90 μL of medium to give final
concentrations of up to 200 mg/L. From the resulting solution,
10 μL was added to each well already containing 200 μL of the
above solution of cells in the medium. After a standard of 3
days incubation at 37 °Cin5%CO2 atmosphere, MTT (0.1 mg
in 20 μL) wasadded, andthecellswere incubatedforadditional
3-4 h. The MTT solution was then removed, and the cells were
dissolved in 200 μL of isopropanol. The absorbance at 550 nm
was measured for 100 μL of the above solution by a Bio-Tek
EL-800 microplate reader spectrophotometer. Each measure-
ment was repeated at least 2 ꢀ 5 times, namely, two repeats per
plate, all repeated five times on different days (10 repeats
altogether). Relative IC50 values were determined by a non-
linear regression of a variable slope (four parameters) model.
Control experiments were also conducted with free ligands and
L4Ti(OiPr)2. was synthesized similarly in quantitative yield
from Ti(OiPr)4 (0.050 g, 0.18 mmol) and L4H2 (0.063 g, 0.18
mmol), and may be recrystallyzed from diethylether (0.041 g,
1
43%). H NMR (CDCl3) δ 6.73 (2H, dd, J = 8.6, 3.4 Hz, Ar),
6.60 (2H, d, J=8.8Hz, Ar), 6.53(2H, d,J= 3.2 Hz, Ar), 5.03 (2H,
sept, J = 6.1 Hz, OCH(CH3)2), 4.63 (2H, d, J = 14.0 Hz, CH2),
3.74 (6H, s, OCH3), 3.06 (2H, d, J = 14.0 Hz, CH2), 2.99 (2H, d,
J = 9.4 Hz, CH2), 2.44 (6H, s, CH3), 1.78 (2H, d, J = 9.4 Hz,
CH2), 1.24 (6H, d, J = 6.1 Hz, OCH(CH3)2, 1.23 (6H, d, J = 6.1
Hz, OCH(CH3)2); 13C NMR (CDCl3) δ 156.2, 151.3, 125.0, 117.7,
115.0, 113.9, 77.6, 64.5, 55.8, 51.9, 47.1, 26.0, 25.7; Anal. Calcd for
C26H40N2O6Ti: C, 59.54; H, 7.69; N, 5.34. Found: C, 59.03; H,
7.19; N, 5.32.
Crystal Data for L5Ti(OiPr)2. C24H32Cl4N2O4Ti, M = 602.22,
˚
monoclinic, a = 8.8097(5), b = 12.7717(8), c = 25.197(2) A, β =
3
˚
97.547(1)°, V = 2810.5(3) A , T = 223(1) K, space group P21/c,
Z = 4, μ(Mo-KR) = 0.717 mm-1, 30700 reflections measured,
6144 unique (Rint = 0.0281). Observed reflections [I > 2σ(I)] =
5809 for which R = 0.0815 and wR2 = 0.1734.
(39) Kim, S. H.; Lee, J.; Kim, D. J.; Moon, J. H.; Yoon, S.; Oh, H. J.; Do,
Y.; Ko, Y. S.; Yim, J. H.; Kim, Y. J. Organomet. Chem. 2009, 694, 3409–
3417.

1038 Inorganic Chemistry, Vol. 50, No. 3, 2011
Peri et al.
3
L6Ti(OiPr)2. was synthesized similarly in quantitative yield
from Ti(OiPr)4 (0.050 g, 0.18 mmol) and L6H2 (0.070 g, 0.18
mmol), and may be recrystallyzed from diethylether (0.041 g,
V = 3777.4(4) A , T = 173(1) K, space group P1, Z = 2,
μ(Mo-KR) = 0.581 mm-1, 42004 reflections measured, 16287 unique
(Rint=0.0225). Observed reflections [I > 2σ(I)]=13890 for which
R=0.0646 and wR2=0.1886. Refinement of the structure revealed
along with the well behaved complex, also three ill resolved and
heavily disordered diethylether fragments. These fragments were
treated with restraints (O-C:1.40 and C-C:1.50\%A) using the
DFIX option of SHELXL-97 and refined isotropically.
˚
1
41%). H NMR (CDCl3) δ 7.09 (2H, d, J = 1.9 Hz, Ar), 6.66
(2H, d, J = 1.6 Hz, Ar), 5.32 (2H, sept, J = 6.1 Hz, OCH(CH3)2),
4.68 (2H, d, J = 13.5 Hz, CH2), 3.10 (2H, d, J = 13.5 Hz, CH2),
2.95 (2H, d, J= 9.6 Hz, CH2), 2.45 (6H, s, CH3), 2.21(6H, s, CH3),
1.79 (2H, d, J = 9.4 Hz, CH2), 1.32 (6H, d, J = 6.0 Hz,
OCH(CH3)2, 1.22 (6H, d, J = 6.2 Hz, OCH(CH3)2); 13C NMR
(CDCl3) δ 155.1, 129.8, 128.5, 126.8, 125.4, 121.3, 78.4, 64.4, 51.6,
47.0, 26.1, 25.6, 20.3; Anal. Calcd for C26H38Cl2N2O4Ti: C, 55.63;
H, 6.82; N, 4.99. Found: C, 55.47; H, 6.92; N, 4.79.
Acknowledgment. We thank Dr. Shmuel Cohen for crystal-
lography. This research received funding from the European
Research Council under the European Community’s Seventh
Framework Programme (FP7/2007-2013)/ERC Grant agree-
ment n° [239603]. The research was also partly supported the
Israel Science Foundation (Grant 124/09) and the Israel Cancer
Research Fund (ICRF).
L7Ti(OiPr)2. was synthesized similarly in quantitative yield
from Ti(OiPr)4 (0.050 g, 0.18 mmol) and L7H2 (0.086 g, 0.18
mmol), and may be recrystallyzed from diethylether (0.054 g,
1
46%). H NMR (CDCl3) δ 7.27 (2H, d, J = 1.5 Hz, Ar), 6.70
(2H, d, J = 1.4 Hz, Ar), 5.36 (2H, sept, J = 6.1 Hz, OCH(CH3)2),
4.69 (2H, d, J = 13.5 Hz, CH2), 3.09 (2H, d, J = 13.6 Hz, CH2),
2.49 (2H, d, J =9.4 Hz, CH2), 2.46 (6H, s, CH3), 2.21 (6H, s, CH3),
1.79 (2H, d, J = 9.4 Hz, CH2), 1.31 (6H, d, J = 6.1 Hz, OCH-
(CH3)2, 1.22 (6H, d, J = 6.2 Hz, OCH(CH3)2); 13C NMR (CDCl3)
δ156.1, 132.8, 129.4, 127.3, 125.1, 111.8, 78.5, 64.5, 51.6, 47.3, 26.1,
25.7, 20.1; Anal. Calcd for C26H38Br2N2O4Ti: C, 48.02; H, 5.89; N,
4.31. Found: C, 48.08; H, 5.59; N, 4.09.
Supporting Information Available: Crystallographic data for
L5Ti(OiPr)2 and Ti3L6 (μ-O)3, a representative 1H NMR spec-
3
trum taken through a hydrolysis study of L1Ti(OiPr)2 with
proton assignment, 1st order fits of Ti-OiPr hydrolysis for
L1-8Ti(OiPr)2, hydrolysis studies of L2-4,8Ti(OiPr)2, 1H
NMR spectra following water addition of L5-7Ti(OiPr)2, and
cytotoxicity plots of L1,4,5,6,8Ti(OiPr)2 toward HT-29 and OV-
CAR-1 cells. This material is available free of charge via the
Internet at http://pubs.acs.org.
Crystal Data for Ti3L6 (μ-O)3. C60H72Cl6N6O9Ti3 (C4H6O)2-
3
3
(H2O)0.5, M = 1532.36, triclinic, a = 12.9257(7), b = 14.3936(8),
˚
c = 22.576(1) A, R = 101.578(1), β = 97.023(1), γ = 110.178(1)°,