← Back
Tuning the Efficacy of Ruthenium(II)-Arene (RAPTA) Antitumor Compounds with Fluorinated Arene Ligands
Organometallics 2009, 28, 5061–5071
DOI: 10.1021/om900345n
5061
Tuning the Efficacy of Ruthenium(II)-Arene (RAPTA) Antitumor
Compounds with Fluorinated Arene Ligands
Anna K. Renfrew, Andrew D. Phillips, Enrico Tapavicza, Rosario Scopelliti,
Ursula Rothlisberger, and Paul J. Dyson*
Institut des Sciences et Ing
enierie Chimiques, Ecole Polytechnique F
ed
erale de Lausanne (EPFL), CH-1015
Lausanne, Switzerland
Received May 1, 2009
A series of compounds of general formula [Ru(η6-fluoroarene)(pta)Cl2] (fluoroarene =C6H5F,
C6H5CF3, and 1,4-C6H4CH3F; pta=1,3,5-triaza-7-phosphatricyclo[3.3.1.1]decane) have been prepared and characterized spectroscopically. Additionally, X-ray diffraction was employed to
characterize two of the complexes and the corresponding precursors, i.e., [Ru(acac)2(η4-cod)] and
Ru(η6-fluoroarene)(η4-cod)] (cod = cycloocta-1,5-diene). The solubility, pKa’s, and the stability
toward hydrolysis of the [Ru(η6-fluoroarene)(pta)Cl2] complexes were studied, and DFT calculations were performed to assist in rationalizing the observed properties at a molecular level. The
cytotoxicities of the pta-based compounds were evaluated in A2780 ovarian cancer cells, and the
observed activities were correlated to the above-mentioned properties. The rate of hydrolysis of the
Ru-Cl bonds in the C6H5CF3 derivative was found to increase significantly at low pH, which
represents a possible method of tumor targeting based on the reduced pH of this particular cellular
environment.
Introduction
Ruthenium-based compounds are rapidly gaining interest
as potential alternatives to platinum-based chemotherapeutic agents, with some examples having been shown to be
effective against cancers not readily treated by cisplatin.1 In
addition, a number of ruthenium-centered complexes have
been found to display a significantly higher degree of selectivity toward cancerous cells than the leading commercially
available platinum drugs, resulting in reduced damage to
healthy tissue. The mechanism of this selectivity is widely
debated, but one strong possibility stems from the ability of
ruthenium complexes to mimic iron in reversible binding to
plasma proteins such as transferrin, which is correlated with
the overexpressed concentration of receptors for this protein
on the surface of cancer cells. This mechanism provides an
effective means for ruthenium complexes to accumulate in
cancer cells.2 Two ruthenium(III)-based drugs, KP10193 and
NAMI-A,4 have been shown to bind to the iron(III)-binding
sites of transferrin.5 Moreover, both compounds have recently completed phase I clinical trials and demonstrate a
high degree of effectiveness against tumors.
Currently, there is considerable interest in ruthenium(II)
compounds as putative anticancer drugs.6 Part of our work
has focused on ruthenium(II)-arene (RAPTA) complexes,
which are based on a Ru(II) center with an η6-coordinated
arene, a monodentate P-bound pta ligand, and two chloride
ligands, Figure 1, the latter being highly labile depending
upon the conditions. In vitro studies on various RAPTA
complexes (differing by the substitution pattern of the arene
ligand) indicate a greater toxicity toward cancer cells than
healthy cells, and the corresponding IC50 values are relatively
high. Moreover, in vivo, studies with [Ru(η6-C6H6)(pta)Cl2],
abbreviated RAPTA-B, and [Ru(η6-p-C6H4MeiPr)(pta)Cl2],
abbreviated RAPTA-C, revealed a high activity toward
metastatic tumors in combination with very low general
toxicity.7
The selectivity of RAPTA-type complexes toward cancer
cells has been proposed to be, at least in part, pH controlled.
Tumors generally operate under hypoxic conditions, deriving a significant portion of their energy from glycolysis,
resulting in the production of lactic acid, which results in
an overall lowering of cellular pH. Tumor cell environments
with a pH as low as 5.5 have been reported, in comparison to
values of pH 7.2 typically associated with healthy cells.8
The notion that RAPTA compounds could display
pH-dependent activity stemmed from the observation that
*Corresponding author. E-mail: paul.dyson@epfl.ch.
(1) (a) Kostova, I. Curr. Med. Chem. 2006, 13, 1085–1107. (b)
Galanski, M.; Arion, V. B.; Jakupec, M. A.; Keppler, B. K. Curr. Pharm.
Des. 2003, 9, 2078–2089.
(2) Dyson, P. J.; Sava, G. Dalton Trans. 2006, 1929–1933.
(3) Smith, C. A.; Sutherland-Smith, A. J.; Keppler, B. K.; Kratz, F.;
Baker, E. N. J. Biol. Inorg. Chem. 1996, 1, 424–431.
(4) Alessio, E.; Mestroni, G.; Bergamo, A.; Sava, G. Curr. Top. Med.
Chem. 2004, 4, 1525–1535.
(5) Groessl, M.; Hartinger, C. G.; Egger, A.; Keppler, B. K. Metal
Ions Biol. Med. 2006, 9, 111–116.
(6) (a) Melchart, M.; Sadler, P. J. In Bioorganometallics; Jaouen, G.,
Ed.; Wiley-VCH: Weinheim, 2006; pp 39-62. (b) Ang, W. H.; Dyson, P. J.
Eur. J. Inorg. Chem. 2006, 20, 4003–4018.
(7) Scolaro, C.; Bergamo, A.; Brescacin, L.; Delfino, R.; Cocchietto,
M.; Laurenczy, G.; Geldbach, T. J.; Sava, G.; Dyson, P. J. J. Med. Chem.
2005, 48, 4161–4171.
(8) Gerweck, L. E. Semin. Radiat. Oncol. 1998, 8, 176–182.
r 2009 American Chemical Society
Published on Web 08/12/2009
pubs.acs.org/Organometallics
5062
Organometallics, Vol. 28, No. 17, 2009
Figure 1. Examples of [Ru(η6-arene)(pta)Cl2] compounds:
(left) RAPTA-B, arene=benzene; (center) RAPTA-T, arene=
toluene; and (right) RAPTA-C, arene=p-cymene.
RAPTA-C induces significant unwinding of supercoiled
DNA at pH values below 7 without exerting an effect on
DNA at physiological pH.9 Indeed, another recent study has
shown that binding of RAPTA-C to potential biomolecular
targets is an order of magnitude higher at pH 6 compared to
pH 7.2.10 Theoretical and experimental studies indicate that
the active and nonactive species of RAPTA-type complexes
are, respectively, the aqua and hydroxo forms derived from
hydrolysis of one of the chloride ligands. So far it has not
been possible to observe the deprotonation reaction by 1H
NMR spectroscopy, as a significant broadening of the
diagnostic arene signals occurs at higher pH; however,
DFT calculations estimate the pKa to be around pH 9 for
RAPTA-B, implying that the dominant species at physiological pH is the aqua complex, which shows greater ligand
labiality, and hence reactivity.11
With the aim of exploiting the pH difference between a
tumor cell and healthy cell environment, DFT calculations
were carried out to design a complex for which the major
species in a healthy cell would be the less reactive hydroxo
compound, with the more active aqua compound dominating in the lower pH environment of a tumor. Accordingly,
strongly electron-withdrawing groups attached to the arene
ring were shown to modulate the equilibrium between the
hydroxo- and aqua-containing complexes. In this respect
fluorine-substituted aromatic groups provide an ideal means
of lowering the pKa without greatly modifying the steric
profile or hydrophilic properties of RAPTA-type complexes.
However, such electron-poor arenes are comparatively weak
ligands, and few examples of ruthenium(II)-fluoroarene
compounds have been reported, the majority being sandwich complexes of general formula [Ru(η5-C5Me5)(η6fluoroarene)]+, for which a general synthetic procedure
has been developed.12 Therefore, we have investigated the
synthesis of several RAPTA complexes with fluoroarene
ligands, viz., [Ru(η6-fluoroarene)(pta)Cl2], and herein describe these compounds and discuss the effects of the role of
the η6-bound fluorous arene ligand on stability, hydrolysis,
pKa, and cell proliferation.
Results and Discussion
The new RAPTA compounds with fluoroarene ligands
shown in Figure 2 were prepared following a multistep
(9) Allardyce, C. S.; Dyson, P. J.; Ellis, D. J.; Heath, S. L. Chem.
Commun. 2001, 1396–1397.
(10) Groessl, M.; Hartinger, C. G.; Dyson, P. J.; Keppler, B. K. J.
Inorg. Biochem. 2008, 102, 1060–1065.
(11) Gossens, C.; Dorcier, A.; Dyson, P. J.; Rothlisberger, U. Organometallics 2007, 26, 3969–3975.
(12) (a) Koelle, U.; Hoernig, A.; Englert, U. Organometallics 1994,
13, 4064–4066. (b) Fang, X.; Watkin, J. G.; Scott, B. L.; John, K. D.; Kabas,
G. J. Organometallics 2002, 21, 2336–2339. (c) Hayashida, T.; Nagashima,
H. Organometallics 2002, 21, 3884–3888.
Renfrew et al.
Figure 2. Structures of new fluoroarene-substituted RAPTA
compounds 4a-4c.
synthesis involving reduction to a Ru(0) species and subsequent oxidation to regenerate the Ru(II)-fluoroarene species. This is in contrast to the previously reported synthesis of
RAPTA complexes, which involves two steps: reduction of
hydrated RuCl3 with the appropriate hexacyclic diene to give
a chloro-bridged arene dimer,13 followed by reaction of the
dimer with 2 equiv of the pta ligand. The diene is usually
obtained through Birch reduction of the corresponding
arene; however, in the case of halogenated arenes, Birch
reduction is dangerous, possibly resulting in the formation of
the highly reactive and potentially explosive arynes. An
alternative procedure involves thermally induced arene
exchange, which has been used to generate η6-C6Me6- or
η6-C6H3(iPr)3-substituted chloro-ruthenium(II) dimers from
the corresponding p-cymene species at 180 °C.13c However,
our attempts to substitute p-cymene with the electron-withdrawing CF3C6H5 arene by thermal or photochemical methods were unsuccessful, as similarly reported by the research
group of Bennett, who failed to displace η6-benzene or
p-cymene with C6H5CF3 from Ru(η6-arene)Cl2{P(nBu)3}
by irradiation with ultraviolet light.13c Consequently, a
seldom used route (Scheme 1) was used to prepare the
dimeric precursors.
The synthesis of 4a-4c commences with [Ru(acac)2(η4cod)] (1), first reported by Wilkinson, for which an improved
synthesis was described by Powell et al. and employed for
this work. Compound 1 is converted to the reactive
ruthenium(0) species [Ru(η6-napthalene)(η4-cod)] using a
modified but established procedure reported by the groups
of Bennett and Vitulli.14 This compound readily undergoes
facile exchange of the 10-electron naphthalene ligand with
the appropriate fluoroarene, in the presence of a small
amount of acetonitrile, to give [Ru(η6-fluoroarene)(η4-cod)]
(fluoroarene = C6H5F 2a, C6H5CF3 2b, 1,4-C6H4CH3F 2c,
and 1,4-C6H4F2 2d).15 In all cases, chromatographic separation was required using slightly basic alumina as the stationary phase and eluting with pentane. Compound 2a and
meta analogues of 2c and 2d, i.e., [Ru(η6-1,3-C6H4CH3F)(η4cod)] and [Ru(η6-1,3-C6H4F2)(η4-cod)], are known compounds.15 While the mono- and difluoro complexes were
successfully prepared by the described method, 1,3,5-trifluorobenzene proved to be too weakly coordinating to
displace the naphthalene ligand, with attempts involving
low-pressure hydrogenation of [Ru(η4-cot)(η4-cod)] (cot=
cyclooctatetraene) also proving unsuccessful.16
(13) (a) Winkhaus, G.; Singer, H. J. Organomet. Chem. 1967, 7, 487–
491. (b) Zelonka, R. A.; Baird, M. C. Can. J. Chem. 1972, 50, 3063–3072.
(c) Bennett, M. A.; Smith, A. K. Dalton Trans. 1974, 233–241.
(14) Bennett, M. A.; Neuman, H.; Thomas, M.; Wang, X.; Pertici, P.;
Salvadori, P.; Vitulli, G. Organometallics 1991, 10, 3237–3245.
(15) Bodes, G. F.; Heinemann, W.; Jobi, G.; Klodwig, J.; Neumann,
S.; Zenneck, U. Eur. J. Inorg. Chem. 2003, 281–292.
(16) Pertici, P.; Vitulli, G.; Lazzaroni, R.; Salvadori, P.; Barili, P. L.
Dalton Trans. 1982, 1019–1022
Article
Organometallics, Vol. 28, No. 17, 2009
5063
Scheme 1. Synthetic Routes Employed for 4a-4da
a
R represents a fluoroarene, i.e., C6H5F 1a-4a, C6H5CF3 1b-4b,
1,4-CH3C6H4F 1c-4c, and 1,4-C6H4F2 1d-4d.
The [Ru(η6-fluoroarene)(η4-cod)] complexes, 2a-2d, are
oxygen sensitive both in solution and in the solid state, but
readily crystallize by evaporation of a saturated pentane
solution. 1H and 13C NMR spectra each show two signals
characteristic of the cyclooctadiene ligand in addition to the
peaks corresponding to the η6-bonded fluoroarene ligand.
The conversion of 2a-2d into chloride-bridged dimers (3a3d) follows a modified literature method involving the addition of 2 equiv of gaseous HCl in pentane, oxidizing the
Ru(0) to Ru(II) with concomitant reduction of the cod ligand
to cyclooctane. The extremely poor solubility of the chlorobridged dimeric species excluded their complete spectroscopic characterization. Nevertheless, the corresponding
[Ru(η6-fluoroarene)(pta)Cl2] complexes are readily prepared
in good yield from the dimers by addition of 2 equiv of pta in
dichloromethane. Complexes 4a-4c are stable to air, and 4c
is soluble in water and polar organic solvents, while 4a and 4b
are only sparingly soluble. Compound 4d could not be
prepared in pure form and therefore was not studied further.
X-ray Diffraction Studies. Single-crystal X-ray diffraction
was used to characterize 4a and 4c and a number of the
intermediate compounds involved in their preparation,
namely, 1 and 2a-d. No single crystals of the dimers were
obtained due to their insolubility in organic solvents. In
contrast, all of the [Ru(η6-fluoroarene)(η4-cod)] complexes,
2a-d, readily formed large and highly crystalline yellow or
orange needles and blocks. The molecular structure of the
parent compound of the series with the naphthalene ligand,
[Ru(acac)2(η4-cod)] (1), has been previously described.17
However, only powder X-ray diffraction data of 1 are
(17) Crocker, M.; Green, M.; Howard, J. A. K.; Norman, N. C.;
Thomas, D. M. Dalton Trans. 1990, 2299–2301.
(18) (a) Potvin, C.; Manoli, J. M.; Dereigne, A.; Pannetier, G. J. Less
Common Met. 1971, 24, 333–334. (b) Potvin, C.; Pannetier, G. J. Less
Common Met. 1970, 22, 91–98.
(19) Bennett, M. A.; Wilkinson, G. Chem. Ind. 1959, 1519.
Figure 3. ORTEP plot of 1 with non-hydrogen atoms represented by thermal parameters with 50% probably ellipsoids.
Selected bond lengths (Å) and angles (deg): Ru(1)-O(1)
2.066(2), Ru(1)-O(2) 2.064(2), Ru(1)-O(3) 2.058(2), Ru(1)O(4) 2.074(2), Ru(1)-C(11) 2.177(2), Ru(1)-C(14) 2.167(2),
Ru(1)-C(15) 2.169(2), Ru(1)-C(18) 2.172(2 O(1)-Ru(1)O(2) 89.84(6), O(3)-Ru(1)-O(4) 89.76(6), O(2)-Ru(1)-O(3)
89.34(6), Ru(1)-O(1)-C(1) 124.95(15), Ru(1)-O(2)-C(4)
125.34(15), Ru(1)-O(3)-C(6) 126.15(15), Ru(1)-O(4)-C(9)
124.52(15), C(11)-Ru(1)-C(18) 36.87(9), C(14)-Ru(1)C(15) 36.73(9), C(11)-Ru(1)-C(14) 81.22(9), C(14)-Ru(1)C(15) 36.73(9).
available,18 despite being first reported in 1959.19 Using
slightly different crystallization procedures from those described previously, we obtained single crystals that gave
different cell parameters and the space group P21/n, instead
of P21/c, in contrast with the powder X-ray data. The
structure of 1 is shown in Figure 3, with selected bond lengths
and angles provided in the caption. The complex is characterized by a six-coordinate, pseudo-octahedral Ru geometry. All of the Ru-O bonds are equivalent in length within
experimental error, as are the two CdC bonds. A search of
the CSD20 reveals that only seven previously reported complexes are known that feature two acteylacetone groups
bound to Ru, along with a ligand (or ligands) with two
CdC bonds. In the case of 1, the CdC bonds are orientated
nearly parallel with one another, as indicated by the
CdC 3 3 3 CdC torsion angle of 7.65(2)°. This contrasts with
acyclic species featuring two ethene groups, i.e., Ru(acac)2(η2-H2CdCH2)2, where the equivalent angle is 76.4(1)°.
In 1 both the internal O-Ru-O bond angles and the external
O(2)-Ru-O(3) angles are equivalent, these parameters being
similar to those observed in other Ru-acac complexes with
coordinated unsaturated hydrocarbon co-ligands.
The structures of 2a-2d are shown in Figure 4, and a
comparison of selected bond parameters is provided in
Table 1 along with the naphthalene-based precursor and
the benzene analogue reported previously.21 Complex 2b
represents the first crystallographically characterized structure to feature a C6H5CF3 ligand η6-bound to a Ru center.
The structure of 2a contains two crystallographically
(20) Allen, F. H. Acta Crystallogr. 2002, B58, 380–388.
(21) Schmid, M.; Ziegler, M. L. Chem. Ber. 1976, 109, 132–138.
5064
Organometallics, Vol. 28, No. 17, 2009
Renfrew et al.
Figure 4. ORTEP plots of 2a-2d with non-hydrogen atoms represented by thermal parameters with 50% probably ellipsoids. The disorder
of the F atom (2a) and CH3 group and F atoms (2c) has been omitted for clarity. Selected bond lengths and angles are given in Table 1.
Table 1. Comparison of Selected Bond Lengths (Å) and Angles (deg) of 2a-2d and Analogues with Naphthalene (C10H8) and
Benzene (C6H6)
2a Ar =
C6H5F a,b
2b Ar =
C6H5CF3
2c Ar =
C6H4(F)CH3
2d Ar =
C6H4F2
Ar = C10H8
Ar = C6H6
centroid(Ar)-Ru
1.760(2)
1.752(2)
1.751(1)
1.749(2)
1.746(1)
1.780(1)
1.750c
Ru-C(dC)
2.156(5)
2.138(5)
2.152(5)
2.140(5)
2.141(5)
2.151(5)
2.157(5)
2.169(5)
2.147(10)
2.155(10)
0.018(7)
0.028(7)
1.431(7)
1.431(7)
2.169(2)
2.161(2)
2.151(1)
2.153(2)
2.082(4)
2.147(3)
2.167(3)
2.116(4)
2.159(3)
2.140(3)
2.157(2)
2.165(2)
2.126(3)
2.131(3)
2.141(3)
2.142(3)
2.134(5)
2.133(5)
2.132(5)
2.148(5)
2.159(4)
2.128(7)
2.155(6)
2.135(6)
2.135(10)
0.018
0.085
0.019
0.016
0.016
1.419(3)
1.431(3)
1.457(5)
1.471(4)
1.428(4)
1.432(4)
1.418(4)
1.414(4)
1.403(11)
1.423(8)
136.31(14)
137.80(14)
85.89(14)
136.90(5)
137.80(5)
85.29(6)
136.83(9)
138.65(9)
84.52(10)
137.09(7)
137.08(7)
85.84(8)
136.58(8)
137.03(9)
86.38(9)
137.68c
136.83c
85.49c
2.732
2.825
2.679
2.825
2.791
2.792
n/a
2.852
n/a
average
range
CdC
centroid(Ar)-Rump(CdC)d
mp(CdC)-Rump(CdC)
(HC)H 3 3 3 C(dC)
a
Structure contains two crystallographically independent molecules in the unit cell. b The arene component of the molecule is disordered over two
positions. c No atomic position esd’s provided in the cif file to calculate the esd of the centroid or midpoint of the CdC bonds. d The abbreviation mp
refers to midpoint or bisection point of the two CdC bonds of the cod group.
independent molecules in the unit cell, each molecule having
a slightly different geometry, with the primary difference
corresponding to the amount of (H2)C-C(H2) single-bond
twisting within the cod ligand. Furthermore, 2a and 2c
feature positional disorder of the F atom over two sites,
which was resolved using a standard method (see Experimental Section).
The distance between the Ru atom and the centroid of the
fluoroarene ligands is considerably shorter than in the
naphthalene complex, which is correlated with the highly labile
nature of this particular arene ligand. The average bond
distances between the Ru center and olefin bond are equal in
all the systems; however, in complex 2c the distance is significantly different, i.e., by 0.085 Å. For 2a-2d the two CdC
bonds of the cod are not coplanar and are twisted in opposing
directions. The greatest amount of twisting, as measured at the
CdC bisection point, is observed in complex 2d [9.2(2)°]. Such
a distortion is probably due to a combination of electronic
effects from the two electron-withdrawing fluorine atoms,
which also exert a slightly greater steric interaction on the
cod group. As a final point, in contrast to the benzene and
naphthalene complexes, and as expected with the introduction
of F atoms, a substantial number of hydrogen-bonding interactions are present, but will not be discussed here.
The structures of 4a and 4c are shown in Figure 5. Both 4a
and 4c comprise a Ru center with a distorted piano-stool
geometry. The rotation of the arene in 4c is such that the
bulkier methyl group does not eclipse the Ru-Cl or Ru-P
bonds. However, in 4c partial eclipsing of the (Ar)C-F bond
with a Ru-Cl bond is observed. Table 2 lists the relevant
metric parameters of 4a and 4c with a comparison of the
previously reported p-cymene-containing structure, RAPTA-C.22 In general, bond parameters of 4c and RAPTA-C
are very similar, with 4a being somewhat different. The
internal structural parameters of the pta ligand are identical
for all of the complexes.
Structure 4c differs from 4a and RAPTA-C in that no
solvate molecules are found in the unit cell. Instead, a tight
packing of the molecules is observed with short intermolecular hydrogen-bonding interactions. An arene-arene πstacking interaction is observed, with the shortest interarene
(22) Allardyce, C. S.; Dyson, P. J.; Ellis, D. J.; Heath, S. L. Chem.
Commun. 2001, 1396–1397.
Article
Organometallics, Vol. 28, No. 17, 2009
5065
Figure 5. ORTEP plots of 4a and 4c with non-hydrogen atoms represented by thermal parameters with 50% probably ellipsoids.
Selected bond lengths and angles are given in Table 2.
Table 2. Selected Bond Lengths (Å) and Angles (deg) of 4a, 4c, and
RAPTA-C
4a
4c
RAPTA-Ca
Ru-Ar(centroid)
1.725(1)
1.705(1)
Ru-Cl
2.434(1)
2.431(1)
2.410(1)
2.411(1)
Ru-P
2.279(1)
2.297(1)
P-C(N)
1.843(2)
1.839(3)
1.840(3)
1.839(2)
1.845(2)
1.846(2)
N-C(N)
1.477(3)
1.481(3)
1.503(3)
1.477(3)
1.472(3)
1.475(3)
Cl-Ru-Cl
87.78(2)
87.88(2)
Cl-Ru-P
84.62(2)
87.86(2)
86.19(2)
82.61(2)
centroid(Ar)-Ru-P
131.20(4)
129.68(3)
centroid(Ar)-Ru-Cl
123.63(4)
127.50(4)
127.09(3)
128.29(3)
Ru-P-C
116.92(8)
119.37(8)
120.49(8)
115.44(7)
119.98(7)
120.82(7)
C-P-C
97.82(12)
98.52(12)
99.31(11)
98.58(9)
98.31(9)
99.45(10)
1.692(4)
1.701(4)
2.412(3)
2.429(3)
2.425(3)
2.425(3)
2.296(2)
2.298(3)
1.840(11)
1.837(10)
1.855(8)
1.827(10)
1.834(11)
1.831(10)
1.470(12)
1.489(12)
1.448(10)
1.474(12)
1.465(13)
1.471(13)
87.25(9)
88.97(9)
87.09(10)
83.42(9)
85.29(10)
82.79(9)
129.60(15)
120.76(15)
126.85(19)
127.86(17)
127.78(18)
127.23(16)
117.3(3)
119.9(3)
119.9(3)
118.0(3)
119.7(3)
119.1(4)
98.4(4)
98.6(5)
98.2(4)
99.9(5)
98.0(4)
97.7(5)
parameter
a
Structure contains two crystallographically independent molecules
within the unit cell. Taken from ref 23.
distance being 3.283(3) Å. A strong intermolecular hydrogen
bond (2.65 Å) between the methyl group and F atom
probably assists in enforcing this stacking arrangement. In
4a a similar arene-arene π-stacking arrangement is observed, except the slippage is greater, and the shortest inter-
arene distance is 3.351(4) Å. Furthermore, the unit cell is
extremely large and a high amount of void space is present.
Within this space is an array of interconnected disordered
water molecules that form a number of intermolecular
hydrogen bonds within the complex, specifically at the ring
fluorine group, chlorine, and the hydrogen atoms of the pta
ligand. Interestingly, the nitrogen centers of the pta are not
involved in hydrogen bonding with the water solvates, which
is generally thought to be responsible for the high degree of
water solubility associated with the pta ligand.23
Characterization of 4a-4c in Aqueous Solution. The behavior of 4a-4c in aqueous solution was carefully monitored
by 1H and 31P NMR spectroscopy. RAPTA-type complexes
have previously been shown to undergo rapid hydrolysis in
water through loss of a chloride ligand (eq 1), to reach an
equilibrium between the dichloro and monochloro complexes.7 The hydrolysis product is observed in the 31P
NMR spectrum by a second peak, ca. 2 ppm downfield of
the starting compound, and the process is reversed through
the addition of sodium chloride.
In contrast to the benzene- and alkyl substituted arene
RAPTA compounds, hydrolysis of 4a-4c is not rapid, and
a new peak corresponding to the hydrolysis species is
visible in the 31P NMR spectrum only after approximately
30 min in solution, with 4a and 4c reaching equilibrium after
2 h and 4b only after ca. 1 day. When the concentration of
Cl- is increased to 100 mM, the 31P NMR signal for the
hydrolysis product is not observed following 3 days in
solution; however, in the case of 4b, and to a lesser extent
4c, a new species appears that exhibits a peak at ca. 1 ppm
from the original peak. The presence of free arene in the 1H
NMR spectra suggests that this species results from the loss
of the arene ligand. While it was not possible to obtain a
crystal structure for 4b, it is not unreasonable to expect the
Ru-arene bond to be relatively weak given the electronwithdrawing nature of the arene ligand. Indeed the Ruarene bond length of the precursor complex 2b is considerably elongated; see Table 1.
Uniquely in the case of 4b, the rate of hydrolysis was found
to increase progressively on the addition of acid with a 5-fold
increase in rate observed on going from pH 5.7 (unbuffered
solution) to 4.7 (Table 3, Figure 6). This rate increase may be
due to more extensive hydrogen bonding between the chloride ligands and protons in solution, thereby activating
the Ru-Cl bond. Consequently, it is conceivable that the
(23) Phillips, A. D.; Gonsalvi, L.; Romerosa, A.; Vizza, F.; Peruzzini,
M. Coord. Chem. Rev. 2004, 248, 955.
5066
Organometallics, Vol. 28, No. 17, 2009
Renfrew et al.
formation of the labile aqua complex would occur more
quickly in a cancerous cell than a healthy cell, also potentially
contributing to drug selectivity.
31
P NMR-based titrations were used to determine the pKa
values for the N-protonation of the pta ligand in 4a-4c (pKa 1,
eq 2, Table 4). It was not possible to observe deprotonation of
the water ligand by NMR (pKa 2, eq 3). The pKa 1 values were
unchanged within experimental error when the measurements
were repeated with a Cl- concentration of 100 mM.
þH2 O, Cl ½Ruðη6 -areneÞðptaÞCl2 s
' ½Ruðη6 -areneÞðptaÞClðH2 OÞ þ
&s
þCl -H2 O
Hþ
þ
6
½Ruðη6 -areneÞðptaÞCl2 &s
s' ½Ruðη -areneÞðpta-HÞCl2
-H þ
Hþ
þ
6
½Ruðη6 -areneÞðptaÞClðOHÞ &s
s' ½Ruðη -areneÞðptaÞClðH2 OÞ
-H þ
ð1Þ
ð2Þ
ð3Þ
In general, the calculations underestimate the experimental
pKa 1 values by 0.12-1.04 units. Most of the alkyl and
fluorinated substituted systems have similar calculated and
experimental values, ranging from 1.9-2.9 and 2.12-3.31,
respectively. However, the calculated pKa values of the
Table 3. kobs for the Hydrolysis of 4b
kobs(s-1)
solution pH
7.05 10-5 ( 3.65 10-6
3.70 10-4 ( 7.51 10-5
5.7
4.7
Figure 6. Decay of absorbance at 355 nm for 4b (pH 5.7 in
unbuffered solution). Data fitted to Abs = A0 exp(-kobst) + A1.
C6H5CF3- and 1,4-FC6H4Me-containing complexes 4b and
4c are significantly reduced (2.0) compared to the other
calculated values. In particular 4b is nearer to the experimentally observed value with a difference of 0.07, accounting
for the error factor associated with the experimental value.
With the exception of compound 4c and RAPTA-C, the
theoretical and experimental values for the first pKa follow
the same trend, namely, decreasing in the order RAPTA-T >
RAPTA-B > RAPTA-F (4a) > RAPTA-CF3 (4b). For
RAPTA-C and 4c, featuring sterically more demanding arene
ligands, the differences between calculated and experimental
values are larger.
Regarding the protonation of the ruthenium-bound OH
ligand (pKa 2), calculations predict remarkably lower values
for the doubly and triply fluorinated compounds (5.1-6.0)
than for the nonfluorinated and singly fluorinated compounds (8.2-9.2). The pKa 2 values follow the trend RAPTA-B>4c > RAPTA-C>4a>RAPTA-T . RAPTA-F2>
4b> RAPTA-F3, decreasing with increasing number of
electron-withdrawing substituents on to η6-arene ligand.
With the exception of 4c, proton affinities (PA) and gas
phase basicities (GB) of the OH group of the fluoroarene
compounds are all lower than the other compounds (GB 230234 vs 237-240 kcal mol-1, respectively) and indicate that the
decrease in pKa 2 is mainly due to electronic effects rather than
solvation effects. Comparing the values of PA and GB for the
pta and the OH groups, the calculations predict an absolute
decrease of the negative electrostatic potential (ESP) induced
by the F-substitution for the OH group, compared to the more
distant N atoms of the pta ligand.
The pKa 2 of 4a and 4c is greater than 7.2, indicating that
similar to known RAPTA compounds, the labile aqua complex would be the dominant species at cellular pH. In contrast,
4b, RAPTA-F2, and RAPTA-F3 have calculated pKa 2 values
below 6.8, indicating that the hydroxyl form would be expected to dominate at cellular pH and the more reactive aqua
form at the reduced pH of cancer tissue, potentially providing
a greater selectivity in that they should be more reactive in the
reduced pH environment of cancer tissue and less reactive in
healthy tissue, where the pH is ca. 7.2.
Cytotoxicity Assays. The cyctotoxicity of 4a-4c was
evaluated in the A2780 human ovarian cancer cell line using
the colormetric MTT assay (Table 6). Complex 4a is less
active than the benchmark compound, RAPTA-C, also
included as a control. In contrast, 4b and 4c are significantly
more cytotoxic than RAPTA-C, and it is noteworthy that the
most active compound, i.e., 4b, corresponds to the one
predicted to have the most relevant pKa 2 value from the
computational study. It is also possible that the more labile
arene ligand in 4b and 4c, compared to other RAPTA
Table 4. Calculated and Experimentally Determined pKa 1 Values with the Corresponding Proton Affinities (PA), Gas Phase Basicities
(GB), and Dipole Moments (μ)
complex
arene
pKa 1 (expt)
pKa 1 (calc)
ΔpKa 1
(pKa(expt) - pKa(calc))
PA (kcal mol-1)
GB (kcal mol-1)
μ (Db)
RAPTA-C
RAPTA-B
RAPTA-T
RAPTA-F (4a)
RAPTA-CF3 (4b)
RAPTA-FT (4c)
RAPTA-F2
RAPTA-F3
1,4-iPrC6H4Me
C6H6
C6H5CH3
C6H5F
C6H5CF3
1,4-FC6H4CH3
1,4-C6H4F2
1,3,5-C6H3F3
3.13 ( 0.02a
3.23 ( 0.06a
3.31 ( 0.03a
3.13 ( 0.04
2.12 ( 0.05
3.04 ( 0.04
-
2.1b
2.8b
2.9
2.5 (2.9)
2.0 (2.6)
2.0 (2.2)
2.30
1.9b
1.03
0.43
0.41
0.63
0.12
1.04
-
222.1b
228.1
228.9
226.5
224.3
227.5
225.7
224.4b
219.3b
220.5
221.3
219.2
216.8
219.8
218.2
216.9b
7.49b
7.41
7.21
7.96
8.91
7.78
7.04
6.65b
a
Data taken from ref 5. b Data taken from ref 8.
Article
Organometallics, Vol. 28, No. 17, 2009
5067
Table 5. Calculated pKa 2 Values with the Corresponding Proton Affinities (PA) and Gas Phase Basicities (GB)a
complex
arene
RAPTA-C
RAPTA-B
RAPTA-T
RAPTA-F (4a)
RAPTA-CF3 (4b)
RAPTA-FT (4c)
RAPTA-F2
RAPTA-F3
1,4-iPrC6H4Me
C6H6
C6H5CH3
C6H5F
C6H5CF3
1,4-FC6H4CH3
1,4-C6H4F2
1,3,5-C6H3F3
pKa 2 (calc)
PA (kcal mol-1)
GB (kcal mol-1)
8.7
9.2b
8.2
8.3 (8.1)
5.5 (6.1)
8.9 (8.2)
6.0
5.1b
246.6
243.9b
244.6
242.3
242.1
245.7
238.2
238.1b
239.4
240.5b
237.8
234.8
234.1
238.5
231.3
230.9b
a
All values were calculated for the conformers with highest solvation energies. The pKa values for the conformer that is most stable in the gas phase are
given in parentheses. b Data taken from ref 8. c Data taken from ref 8.
Table 6. IC50 Values of 4a-4c and RAPTA-C in A2780 Cells
complex
IC50 (μM)
RAPTA-C
RAPTA-F (4a)
RAPTA-CF3 (4b)
RAPTA-FT (4c)
353
507
38
78
compounds, contributes to their greater cytotoxicity. Indeed, it has been shown that, following binding to a model
oligonucleotide, RAPTA derivatives lose the arene ligand,24
and such a process is likely to result in increased distortions
to the oligonucleotide structure due to increased coordination demands of the complex. It should be noted that while
RAPTA-C is scarcely toxic, it is highly effective in vivo,
displaying good activity against both metastatic7 and primary tumors,25 albeit at high doses. Thus while the greater
cytotoxicity of 4b and 4c may lead to activity against
different tumors, they are likely to be applicable at much
lower doses than RAPTA-C, which is very interesting from a
pharmacological point of view.
of this type could react with biologically relevant targets.28 In
this paper a series of RAPTA-type complexes with fluorosubstituted η6-arene ligands have been prepared and characterized using several techniques. In particular the hydrolysis behavior has been compared with other previously
reported RAPTA compounds. pKa values for the protonation of the pta ligand were predicted by DFT with reasonable accuracy for complexes containing mono F-substituted
arene ligands; however, the method was less effective for
di-F-substituted arenes and the trifluoromethyl derivative.
The electron-withdrawing fluoroarene ligands modulate the
pKa’s of the complexes, strongly influencing their behavior in
solution and leading to comparatively low IC50 values.
Moreover, the pH-dependent hydrolysis of 4b may prove
to be an effective “built-in” mechanism for selectively targeting cancerous tissue while sparing healthy tissue due to their
intrinsic differences in pH. It is not unreasonable to assume
that 4b would hydrolyze only in the reduced pH environment
of a tumor, thus becoming active and able to induce cell
death. In healthy tissue 4b should be considerably less labile
and therefore less toxic.
Conclusions
Experimental Section
Rationally designed organometallic compounds are attracting increasing attention as potential anticancer drugs.26
In this context organoruthenium(II) compounds show much
promise,27 following early studies showing that compounds
[Ru(η4-C8H12)Cl2]n, [Ru(η4-C8H12)(η6-napthalene)], and pta
were prepared according to literature methods.29,14,30 Compounds were prepared under an inert atmosphere using dried
and degassed solvents. 1H, 13C, and 31P NMR spectra were
recorded at 400 MHz on a Bruker Avance DPX spectrometer at
room temperature. 19F NMR spectra were recorded at 200 MHz
on a Bruker Avance DPX spectrometer at room temperature.
ESI-MS of the complexes were obtained on a Thermo Finnigan
LCQ Deca XP Plus quadrupole ion trap instrument set in
positive mode (solvent: methanol; flow rate: 5 μL/min; spray
voltage: 5 kV; capillary temperature: 100 °C; capillary voltage:
20 V), as described previously.31
Synthesis of [Ru(η4-C8H12)(acac)2], 1. A modified synthesis of
the one reported by Powell was used.32 To a two-neck flask with
reflux condenser and gas-inlet was added [Ru(η4-C8H12)Cl2]n
(2.72 g, 10 mmol) and anhydrous Na2CO3 (10 g, Fluka). The
flask and its contents were evacuated and flushed with N2. A
solution of 2,4-pentanedione (2.925 g, Aldrich, vacuum distilled) in DMF (50 mL, Acros, degassed and dried over 4 Å
molecular sieves) was added to the reaction flask via cannula.
The resulting orange solution was heated at 150 °C for 30 min,
and the solution was filtered in air while hot. Additional DMF
(15 mL) was added to wash the flask and frit. Doubly distilled
water (135 mL) was added to the filtrate, over a period of 1 h,
causing the precipitation of an orange-yellow solid. The product
(24) Scolaro, C.; Geldbach, T. J.; Rochat, S.; Dorcier, A.; Gossens,
C.; Bergamo, A.; Cocchietto, M.; Tavernelli, I.; Sava, G.; Rothlisberger,
U.; Dyson, P. J. Organometallics 2006, 25, 756–765.
(25) Chatterjee, S.; Kundu, S.; Bhattacharyya, A.; Hartinger, C. G.;
Dyson, P. J. J. Biol. Inorg. Chem. 2008, 13, 1149–1155.
(26) For example see: (a) Nguyen, A.; Top, S.; Pigeon, P.; Vessieres,
A.; Hillard, E. A.; Plamont, M.; Huche, M.; Rigamonti, C.; Jaouen, G.
Chem. Eur. J. 2009, 15, 684–696. (b) Top, S.; Thibaudeau, C.; Vessieres, A.;
Brule, E.; Le Bideau, F.; Joerger, J.; Plamont, M.; Samreth, S.; Edgar, A.;
Marrot, J.; Herson, P.; Jaouen, G. Organometallics 2009, 28 (5), 1414–1424.
(c) Strohfeldt, K.; Tacke, M. Chem. Soc. Rev. 2008, 37, 1174–1187. (d)
Hogan, M.; Claffey, J.; Pampillon, C.; Tacke, M. Med. Chem. 2008, 4, 91–
99. (e) Kirin, S. I.; Ott, I.; Gust, R.; Mier, W.; Weyhermueller, T.; MetzlerNolte, N. Angew. Chem. 2008, 47, 955–959. (f) Gross, A.; Metzler-Nolte, N.
J. Organomet. Chem. 2009, 694, 1185–1188. (g) Zobi, F.; Blacque, O.;
Sigel, R. K. O.; Alberto, R. Inorg. Chem. 2007, 46, 10458–10460. (h) Xavier,
C.; Giannini, C.; Dall'Angelo, S.; Gano, L.; Maiorana, S.; Alberto, R.; Santos, I.
J. Biol. Inorg. Chem. 2008, 13, 1335–1344. (i) Hartinger, C. G.; Dyson, P. J.
Chem. Soc. Rev. 2009, 38, 391–401.
(27) (a) Romerosa, A.; Saoud, M.; Campos-Malpartida, T.; Lidrissi,
C.; Serrano-Ruiz, M.; Peruzzini, M.; Garrido, J. A.; Garcia-Maroto, F.
Eur. J. Inorg. Chem. 2007, 18, 2803–2812. (b) Schmid, W. F.; John, R. O.;
Arion, V. B.; Jakupec, M. A.; Keppler, B. K. Organometallics 2007, 26,
6643–6652. (c) Habtemariam, A.; Melchart, M.; Fernandez, R.; Parsons, S.;
Oswald, I. D. H.; Parkin, A.; Fabbiani, F. P. A.; Davidson, J. E.; Dawson, A.;
Aird, R. E.; Jodrell, D. I.; Sadler, P. J. J. Med. Chem. 2006, 49, 6858–6868.
(28) (a) Fish, R. H. Coord. Chem. Rev. 1999, 185-186, 569–584. (b)
Chen, H.; Maestre, M. F.; Fish, R. H. J. Am. Chem. Soc. 1995, 117, 3631–
3632.
(29) Albers, M. O.; Ashworth, T. V.; Oosthuizen, H. E.; Singleton, E.
Inorg. Synth. 1989, 26, 249–258.
(30) Daigle, D. J. Inorg. Synth. 1998, 32, 40–45.
(31) Dyson, P. J.; McIndoe, J. S. Inorg. Chim. Acta 2003, 354, 68–74.
(32) Powell, P. J. Organomet. Chem. 1974, 65, 89–92.
5068
Organometallics, Vol. 28, No. 17, 2009
was collected on a frit, washed with diethyl ether (250 mL), and
then dried under vacuum overnight. Yield: 3.604 g (88%).
Crystals suitable for X-ray diffraction were obtained by slow
evaporation of an acetone solution of 1.
1
H NMR (CD2Cl2) δ (ppm): 1.91 (s, 6H, CH3COCH), 1.941.99 (m, 2H, C8H12 CH2), 2.06-2.12 (m, 2H, C8H12 CH2), 2.16
(s, 6H, CH3COCH), 2.35-2.44 (m, 4H, C8H12 CH2), 4.03 (m,
2H, C8H12 CH), 4.14 (m, 2H, C8H12 CH), 5.34 (s, 1H,
CH3COCH), 5.36 (s, 1H, CH3COCH). 13C NMR (CD2Cl2) δ
(ppm): 27.19 (s, C8H12 CH2), 27.90 (s, C8H12 CH2), 28.29 (s,
CH3COCH), 30.12 (s, CH3COCH), 88.56 (s, C8H12 CH), 91.76
(s, C8H12 CH), 98.37 (s, CH3COCH), 185.75 (s, CH3COCH),
186.50 (s, CH3COCH).
General Method for the Preparation of 2a-2d. Compounds
2a-2d were prepared using the method described by Bodes
et al.15 The complex [Ru(η6-C10H8)(η4-C8H12)] (1.083 g,
3.2 mM) and the appropriate arene (20 mL, Acros, degassed
and dried over 4 Å molecular sieves) were stirred in THF
(10 mL, dried and degassed) and MeCN (0.833 g, extra dry
quality from Acros) for 7-12 days. The brown solution became
slightly lighter in color over time. The solvents were removed
under vacuum, and the remaining brown reside was extracted
with dried and degassed pentane (25 mL), affording a clear
yellow-brown solution. Using a Brockman grade III oxygenfree Al2O3 packed column (25 cm1 cm) with filter adaptor, the
solution was eluted with pentane (80 mL), yielding a bright
yellow solution. Volatiles were removed under vacuum,
affording a crystalline yellow product, which was dried under
vacuum for 2 h. Crystals suitable for X-ray diffraction studies
were obtained by slow evaporation of a saturated pentane
solution.
[Ru(η6-C6H5F)(η4-C8H12)], 2a. Yield: 579 mg (66%), yellow
powder. 1H NMR (C6D6) δ (ppm): 2.38 (m, 8H, C8H12 CH2),
3.61 (m, 4H, C8H12 CH), 4.21 (td, 3JHH=5.5 Hz, 3JHF=3.8 Hz,
1H, C6H5F p-CH), 5.15 (dd, 3JHH =5.5 Hz, 3JHF=2.7 Hz, 2H,
C6H5F o-CH), 4.74 (dd, 3JHH = 5.5 Hz, 3JHF = 2.7 Hz, 2H,
C6H5F m-CH). 13C NMR (C6D6) δ (ppm): 31.69 (s, C8H12
CH2), 61.35, (s, C8H12 CH), 80.79, (s, C6H5F o-CH), 81.16 (s,
C6H5F m-CH), 86.13 (s, C6H5F p-CH). 19F NMR (C6D6) δ
(ppm): -100.01 (F-C6H5).
[Ru(η6-C6H5CF3)(η4-C8H12)], 2b . Yield: 932 mg (82%), yellow
powder. 1H NMR (C6D6) δ (ppm): 2.255-2.397 (m, 8H, C8H12
CH2), 3.672 (m, 4H, C8H12 CH), 4.628 (m, 3JHH =5.80, 5.92 Hz,
2H, C6H5CF3 m-CH), 4.909 (t, 3JHH = 5.80 Hz, 1H, C6H5CF3
p-CH), 5.101 (d, 3JHH = 5.92 Hz, C6H5CF3 o-CH). 13C NMR
(25 °C, 100.1 MHz, C6D6) δ (ppm): 33.69 (s, C8H12 CH2), 63.35
(s, C8H12 CH), 82.79 (s, C6H5CF3 p-CH), 83.36 (q, 3JCF=43 Hz,
C6H5CF3 o-CH), 86.18 (q, 1JCF =272 Hz, C6H5CF3). 19F NMR
(25 °C, 188.1 MHz, C6D6) δ (ppm): -60.3 (s, 1JCF = 272 Hz,
C6H5CF3).
[Ru(η6-1,4-CH3C6H4F)(η4-C8H12)], 2c . Yield: 73 mg (85%),
yellow powder. 1H NMR (C6D6) δ (ppm): 1.62 (s, 3H,
CH3C6H5CF), 2.43 (m, 8H, C8H12 CH2), 3.50 (m, 4H, C8H12
CH), 4.33 (dd, 3JHH =2.4 Hz, 4JFH =2.8 Hz, 2H, CH3C6H4F
o-CH), 5.06 (dd, 3JHH =2.4 Hz, 3JFH =2.8 Hz, 2H CH3C6H4F
m-CH). 13C NMR (C6D6) δ (ppm): 17.59 (d, 5JFC=7.6 Hz,
CH3C6H4F), 34.06 (s, C8H12 CH2), 64.32 (s, C8H12 CH), 73.77
(s, CH3C6H4F o-CH), 82.82 (d, 3JFC = 25.2 Hz, CH3C6H4F
m-CH).
[Ru (η6-1,4-C6H4F2)(η4-C8H12)], 2d . Yield: 243 mg (43%),
yellow powder. 1H NMR (C6D6) δ (ppm): 2.26-2.44 (m, 8H,
C8H12 CH2), 3.66 (m, 4H, C8H12 CH), 5.31 (d, 3JHF =5.9 Hz,
1,4-F2-C6H4). 13C NMR (C6D6) δ (ppm): 35.23 (s, C8H12 CH2),
64.72 (s, C8H12 CH), 82.79, 132.6 (s, 1,4-F2-C6H4 CH), (s, 1,4F2-C6H4 CF). 19F NMR (C6D6) δ (ppm): -138.2 (1,4-F2-C6H4).
General Method for the Preparation of 3a-3d. To a 50 mL
Schlenk flask was added 2 (0.80 mM) under N2 and dissolved
with pentane (15 mL, dried), affording a clear yellow solution.
The flask was connected with a Y-tube to a N2-vacuum line and
a HCl gas cylinder with inlet adaptor. After five cycles of
Renfrew et al.
purging with N2 the HCl cylinder was opened, and after 2 min
the solution turned red and a precipitate formed. The mixture
was stirred for 1 h under N2 and then filtered (in air) and washed
with pentane (100 mL), then Et2O (25 mL). Finally, CH2Cl2
(5 mL, HPLC grade) was added, then Et2O (10 mL) such that the
resulting filtrate was coloress. The orange-red products were
dried under vacuum overnight.
[Ru(η6-C6H5F)Cl2]2, 3a. Yield: 540 mg (54%), red powder. 1H
(d6-DMSO) δ (ppm): 5.58 (t, 3JHH=5.6 Hz, 1H, C6H5F p-CH),
6.12 (dt, 3JHH = 5.6, 5.6 Hz, 2H, C6H5F m-CH), 6.34 (d, 2H,
3
JHH =5.6 Hz, C6H5F o-CH). 13C NMR (d6-DMSO) δ (ppm):
70.539 (s, C6H5F o-CH), 78.59 (s, C6H5F m-CH), 84.37 (s,
C6H5F p-CH), 145.88 (s, C6H5F CF). 19F NMR (d6-DMSO) δ
(ppm): -128 (C6H5F).
[Ru(η6-C6H5CF3)Cl2]2, 3b. Yield: 189 mg (74.6%), orange
powder. 1H NMR (d6-DMSO) δ (ppm): 5.91 (dd, 3JHH=5.6, 5.8
Hz, 2H, C6H5CF3 m-CH), 6.04 (t, 3JHH=5.6 Hz, 1H, C6H5CF3
p-CH), 6.22 (d, 3JHH=5.8 Hz, 2H, C6H5CF3 o-CH). 19F NMR
(d6-DMSO) δ (ppm): -61.3 (C6H5CF3 CF3).
[Ru(η6-1,4-CH3C6H4F)Cl2]2, 3c. Yield: 81 mg (95%), orange
powder.
[Ru(η6-1,4-C6H4F2)Cl2]2, 3d. Yield: 121 mg (82.9%), red
powder. 1H NMR (d6-DMSO) δ (ppm): 5.65 (m, 3JHF = 4.8
Hz, 1,4-F2-C6H4). 19F NMR (d6-DMSO) δ (ppm): -102.4 (1,4F2-C6H4).
Synthesis of [Ru(η6-C6H5F)Cl2(pta)], 4a. [Ru(η6-C6H5F)Cl2]2
(40 mg, 0.075 mM) and pta (23 mg, 0.15 mM) in CH2Cl2 were
stirred at RT for 4 h. The product was precipitated by addition
of diethyl ether, and the precipitate filtered and washed with
diethyl ether (10 mL) and pentane (10 mL). A concentrated
solution in H2O was stored at 4 °C to give crystals suitable for
X-ray diffraction. Yield: 32 mg (92%) of pale orange powder.
1
H NMR (D2O) δ (ppm): 4.23 (s, 6H, pta), 4.47 (s, 6H, pta),
5.27 (t, 3JHH = 5.5 Hz, 1H, C6H5F p-CH), 5.84 (d, 3JHH =5.6
Hz, 2H, C6H5F o-CH), 6.19 (dd, 3JHH=5.5, 5.6 Hz, 2H, C6H5F
m-CH). 31P{1H}NMR (D2O) δ (ppm): -29.66 (ptaRuCl2), 28.40 (ptaRuClH2O). ESI-MS (0.1 M NaCl) m/z 389.6 [Ru(η6C6H5F)Cl(pta)]þ (56%), m/z 424.9 [Ru(η6-C6H5F)Cl2(pta)]
(100%), m/z=448.0 [Ru(η6-C6H5F)Cl2(pta)]Naþ (100%).
Synthesis of [Ru(η6-C6H5CF3)Cl2(pta)], 4b. [Ru(η6-C6H5CF3)Cl2]2 (58 mg, 0.09 mM) and pta (28 mg, 0.18 mM) were
stirred in CH2Cl2. The product formed as a precipitate following
several minutes of stirring. The reaction was continued for a
further 4 h, and the product was isolated by filtration and then
washed with diethyl ether (10 mL) and pentane (10 mL). Yield:
75 mg (92%), orange powder.
1
H NMR (D2O) δ (ppm): 4.234 (s, 6H, pta) 4.47 (s, 6H, pta),
5.82 m, 3JHH=5.6, 6.2 Hz, 2H, C6H5CF3 m-CH), 5.99 (t, 3JHH=
5.6, 1H, C6H5CF3 p-CH), 6.54 (d, 3JHH=6.2 Hz, 2H, C6H5CF3
o-CH). 31P{1H} NMR (D2O) δ (ppm): -30.64. ESI-MS (H2O)
m/z 474.3 [Ru(η6-C6H5CF3)Cl2(pta-H)]þ (100%).
Synthesis of [Ru(η6-1,4-CH3C6H4F)Cl2(pta)], 4c. [Ru(η6-1,4CH3C6H4F)Cl2]2 (50 mg, 0.09 mM) and pta (28 mg, 0.18 mM) in
CH2Cl2 were stirred at RT for 4 h. The product was precipitated
by addition of diethyl ether, and the precipitate was filtered and
washed with diethyl ether (10 mL) and pentane (10 mL). Vapor
diffusion of pentane into a chloroform solution gave crystals
suitable for X-ray diffraction. Yield: 68 mg (88%), orange
powder.
1
H NMR (D2O) δ (ppm): 1.64 (s, 3H, CH3C6H4F), 3.74 (s, 6H,
pta), 3.94 (s, 6H, pta) 4.87 (d, 3JHH=5.2 Hz, 2H, CH3C6H4F oCH), 4.96 (d, 3JHH =5.2 Hz, 2H, CH3C6H4F m-CH). 31P{1H}NMR (D2O) δ (ppm): -32.20. ESI-MS (0.1 M NaCl):
m/z 111.0 C7H7F (90%), m/z 317.7 [RuCl(H2O)(pta)]þNaþ
(40%), m/z 403.9 [Ru(η6-1,4-CH3C6H4F)(OH)(pta)]þ (100%),
m/z 474.3 [Ru(η6-1,4-CH3C6H4F)Cl2(pta)] (70%).
Determination of pKa Values. The pH values of NMR samples
in D2O were measured at 298 K using a 713 pH meter
(Metrohm) equipped with an electrode, which was calibrated
with buffer solutions at pH values of 4, 7, and 9. The pH values
Article
Organometallics, Vol. 28, No. 17, 2009
were adjusted with dilute HCl and NaOH. The pH titration
curves were fitted to the Henderson-Hasselbach equation using
the program Micromath Scientist (Micromath Scientist Software Inc.) with the assmption that the observed chemical shifts
are weighted averages according to the populations of the
protonated and deprotonated species. The resonance frequencies change smoothly with pH between the chemical shifts of the
charged form HAþ, stable in acidic solution, and those of the
neutral, deprotonated form A, which is present at high pH. At
any pH, the observed chemical shift is a weighted average of the
two extreme values δ(HAþ) and δ(A).
δav ¼
δðHAþ Þ½HAþ þδðAÞ½Aþ
½HAþ þ½Aþ
ð4Þ
The midpoint of the titration occurs when concentration of
the acid and its conjugate base are equal: [HAþ]=[A], that is,
when the pH equals the pKa of the compound. The pH at the
midpoint of the curve is corrected by subtracting 0.44 from the
pD values, as measurements were carried out in D2O.33
DFT Calculations. Calculated pKa values for the reaction in
water, BHþaq f Baq þ Hþaq, were computed using
pK a ¼ ½GðHþ Þgas þΔGs ðHþ ÞþGðBÞgas
þΔGs ðBÞ GðBHþ Þgas -ΔGs ðBHþ Þ=2:303RT
ð5Þ
where G(X)gas denotes the free energy of species X (X=Hþ,
B, BHþ) in the gas phase. G(X)gas is accessible by standard
quantum chemical calculations. The computation of 4Gs(X), the change in free energy due to solvation of species X,
can be performed by using a continuum electrostatic model
for the solvent. For derivation of equation 5, the protonation
reaction in solution was decomposed into the analogous
reaction in the gas phase (BHþgas f Bgas þ Hþgas) as well
as into the single solvation reactions of the involved species
(X f Xaq). For more details see Liptak et al.34 Using
constant values for G(Hþ)34,35 and 4Gs(H),36 eq 5 can be
simplified into
pK a ¼ ½GðBÞgas - GðBHþ Þgas þ ΔGs ðBÞ þ ΔGs ðBHþ Þ
- 269:0=1:3644
ð6Þ
The required quantities were computed at the DFT level
employing the B3LYP exchange-correlation functional36,11 as
implemented in the Gaussian03 computer code.37 Calculations
were performed either in the gas phase or using Barone and
Cossi’s polarizable conducter model for the solvent (C-PCM)38
to compute 4Gs(X).
The zero-point energies, thermal corrections, and entropies
obtained from an analytical frequency analysis have been used
to convert the internal energies to Gibbs energies at 298.15 K
and 1 atm. The same procedure and computational parameters
that have been previously used to compute pKa values of the
same class of ruthenium compounds were employed.11 C-PCM
calculations were performed as single points on the gas phase
(33) Mikkelesen, K.; Nielsen, S. O. J. Phys. Chem. 1960, 64, 632–637.
(34) Liptak, M. D.; Gross, K. C.; Seybold, P. G.; Feldgus, S.; Shields,
G. C. J. Am. Chem. Soc. 2002, 124, 6421–6427.
(35) McQuarrie, D. M. Statistical Mechanics; Harper and Row:
New York, 1970.
(36) Becke, A. D. J. Chem. Phys. 1993, 98, 5648–5652.
(37) Gaussian 03, Revision B.03; Gaussian, Inc.: Wallingford, CT,
2004.
(38) (a) Barone, V.; Cossi, M. J. Phys. Chem. A 1998, 102, 1995–2001.
(b) Cossi, M.; Rega, N.; Scalmani, G.; Barone, V. J. Comput. Chem. 2003,
24, 669–681.
5069
optimized geometries since this has been shown to give better
results than reoptimization of the geometries in the presence of
the solvent model.39 Compounds 4a-4c have been found to be
stable in different orientations of the arene ligand with respect to
the rest of the molecule. The different conformers exhibit large
variations in the dipole moments and the solution enthalpies.
Only the pKa values for the conformers that exhibit the largest
solution enthalpies according to the CPCM calculations are
reported. In the case of 4a and 4c, the conformers with the
highest solvation energies are consistent with the determined
crystal structures. Water was modeled with a dielectric constant
of ε = 78.39. For other C-PCM parameters the default values
suggested by Gaussian03 were taken.
A mixed basis set using the quasirelativistic Stuttgart/Dresden semicore SDD-ECP40 with a (8s7p6d)/[6s5p3d]-GTO tripleζ valence basis set on ruthenium and 6-31þG(d) on the remaining atoms was used. This method has been shown to yield
accurate structures and energies.41
X-ray Diffraction Studies. Suitable single crystals were selected and manipulated in a perfluoro-polyalkyl ether oil matrix,
and for compounds 2a-2d all operations were performed under
a protective blanket of N2. The crystals were mounted on the end
of a glass fiber attached to a metal pin fixed to a goniometer
head, which was placed in the Euler cradle, while maintaining a
cold blanket of N2 gas. The data for 2a, 2b, and 2d were collected
on a Nonius KappaCDD diffractometer equipped with a Bruker-Apex II CCD area detector and an Enraf-Nonius FR590
X-ray generator. For 1, 4a, and 4c, an Oxford-Kuma Kappa
diffractometer with a Sapphire CCD area detector and for 2c, a
Marresearch mardtb desktop goniostat with a Mar345 image
plate were employed. All instruments utilize a graphite-monochromatic Mo KR radiation source with λ = 0.71073 Å. The
crystals were kept under a 140 or 100 K gaseous flow of N2
during data collection. For 2a, 2b, and 2d the unit cell and
orientation matrix was determined by indexing reflections measured from phi-chi scans and analyzed with the program DIRAX,42 while for the remaining structures, the unit cells were
determined from the entire data set using CrysAlis RED.43 All
data sets are based on collecting reflections using an optimized
scanning strategy utilizing the programs CollectCCD,44 CrysAlis CCD, or autoMar.45 After data integration with either
EvalCCD46 or CrysAlis RED, a multiscan absorption correction based on a semiempirical method was applied using the
SADABS,47 ABSPACK (a subprogram of the CrysAlis RED).
Space group determination was performed with the XPREP
program.48 A structure solution based on the direct-method
algorithm was employed with SHELXS-97.49 Afterward, anisotropic refinement of all non-hydrogen atoms was completed
based on a least-squares full-matrix method against F2
(39) Halvani, S.; Noorbala, M. R. THEOCHEM 2004, 711, 13–
18.
(40) Andrae, D.; Haeussermann, U.; Dolg, M.; Stoll, H.; Preuss, H.
Theor. Chim. Acta 1990, 77, 123–41.
(41) (a) Dorcier, A.; Dyson, P. J.; Gossens, C.; Rothlisberger, U.;
Scopelliti, R.; Tavernelli, I. Organometallics 2005, 24, 2114–2123. (b)
Kallies, B.; Mitzner, R. J. Phys. Chem. B 1997, 101, 2959–2967. (c) De
Abreu, H. A.; De Almeida, W. B.; Duarte, H. A. Chem. Phys. Lett. 2004,
383, 47–52.
(42) Duisenberg, A. J. M. J. Appl. Crystallogr. 1992, 25, 92–96.
(43) CrysAlis CCD and CrysAlis RED, version 1.71; Oxford Diffraction
Ltd: Abingdon, 2006.
(44) COLLECT: data collection software; Bruker AXIS B.V.: Delft,
1999.
(45) Automar, version 1.8; Marresearch G.m.b.H.: Norderstedt, 2005.
(46) Duisenberg, A. J. M.; Kroon-Batenburg, L. M. J.; Schreurs,
A. M. M. J. Appl. Crystallogr. 2003, 36, 220–229.
(47) Sheldrick, G. M. SADABS: Area detector absorption and other
corrections, version 2.06; Bruker-AXS: Madison, 2003.
(48) XPREP: Reciprocal Space Exploration, version 6.14; BrukerAXS: Madison, 2003.
(49) Sheldrick, G. M. SHELXS-97, Program for Crystal Structure
Solution; University G€ottingen: Germany, 1997.
5070
Organometallics, Vol. 28, No. 17, 2009
Renfrew et al.
Table 7. Selected Crystallographic Data for 1, 2a, 2b, and 2c
parameter
1
2a
2b
2c
)
)
formula
C18H26O4Ru
C14H17F1Ru
C15H17F3Ru
C15H19F1Ru
407.46
305.35
355.36
319.37
fw (g mol-1)
cryst syst
monoclinic
monoclinic
orthorhombic
monoclinic
C2/c
P212121
P21/c
space group
P21/n
a (Å)
7.7873(3)
23.552(5)
8.2277(16)
12.377(3)
b (Å)
16.0694(4)
8.1761(16)
11.069(7)
8.3601(10)
c (Å)
14.4680(4)
24.319(5)
14.291(3)
12.819(3)
R (deg)
90
90
90
90
β (deg)
104.012(3)
101.82(3)
90
109.621(18)
γ (deg)
90
90
90
90
3
1756.61(10)
4583.5(16)
1301.5(4)
1249.4(5)
volume (Å )
Z
4
16
4
4
1.541
1.770
1.814
1.698
Dcalc (g cm-3)
0.909
1.349
1.222
1.241
μ (mm-1)
F(000)
840
2464
712
648
temp (K)
140(2)
100(2)
100(2)
140(2)
measd reflns
10 241
44 739
24977
7274
unique reflns
2986
4025
2286
2197
theta range (deg)/completeness (%)
2.90 to 25.00/96.5
3.42 to 25.00/99.8
3.39 to 25.00/99.7
2.96 to 24.99/99.9
no. of data/ params/restraints
2986/212/0
4025/309/66
2286/172/0
2197/164/0
1.016
1.505
1.079
1.152
GooFa
Rb [I > 2σ(I)]
0.0249
0.351
0.0126
0.0280
0.0652
0.836
0.0296
0.0724
wR2b (all data)
0.959/-0.427
0.789/-1.096
0.198/-0.236
1.035 and -0.742
largest diff peak/hole (e Å-3)
P
P
P
a
1/2
b
2
2 2
GooF
Fo| - |Fc / |Fo|,
Pis defined as { P[w(Fo - Fc ) ]/(n - p)} where n is the number of data and p is the number of parameters refined. R =
wR2 = { [w(F2o - F2c )2]/ [w(F2o)2]}1/2.
Table 8. Selected Crystallographic Data for 2d, 4a, and 4c
parameter
2d
4a
4c
data using SHELXL-97.50 Hydrogen atoms were added in
geometrically calculated positions and refined as a riding model
using a scaled thermal parameter to the connecting atom. In 2a
and 2c, positional disordering of the fluorine and methyl group
(2c only) was observed and resolved by splitting the atoms over
two positions and allowing the total occupancy of the disordered groups to freely refine. For structure 4a, a number of
disordered water molecules were observed. The total occupancy
for all of the water molecules was restricted to a total of 2.5 units,
which is in accordance with the number of electrons associated
(50) Sheldrick, G. M. SHELXL-97, Program for Crystal Structure
Refinement; University G€ottingen: Germany, 1997.
)
)
formula
C14H16F2Ru
C12H22Cl2F1N3O2.50P1Ru
C13H19Cl2F1N3P1Ru
323.34
470.27
439.25
fw (g mol-1)
cryst syst
orthorhombic
monoclinic
triclinic
P21/c
P1
space group
P212121
a (Å)
6.4319(5)
9.4040(2)
7.0518(3)
b (Å)
7.6925(7)
15.7312(3)
10.8932(5)
c (Å)
23.9146(19)
11.9961(4)
10.9145(5)
R (deg)
90
90
110.878(4)
β (deg)
90
105.972(3)
94.109(3)
γ (deg)
90
90
94.542(3)
1183.23(19)
1706.15(8)
776.40(6)
volume (Å3)
Z
4
4
2
1.815
1.831
1.879
Dcalc (g cm-3)
1.323
1.348
1.463
μ (mm-1)
F(000)
648
948
440
temp (K)
100(2)
140(2)
140(2)
measd reflns
15694
9689
4639
unique reflns
1078
2671
2346
theta range (deg)/completeness (%)
3.41 to 25.00/99.6
2.77 to 25.00/88.8
3.32 to 25.00/86.4
no. of data/ params/restraints
2078/156/0
2671/237/4
2346/191/0
1.097
1.054
1.053
GooFa
0.0200
0.0210
0.0150
Rb [I > 2σ(I)]
0.0489
0.0616
0.0392
wR2b (all data)
-3
0.334/-1.263
0.695/-0.546
0.243/-0.307
largest diff peak/hole (e Å )
P
P
P
a
GooF
as { P[w(F2o - F2c )2]/ (n - p)}1/2 where n is the number of data and p is the number of parameters refined. b R =
Fo| - |Fc / |Fo|,
Pis defined
2
2 2
2 2 1/2
wR2 = { [w(Fo - Fc ) ]/ [w(Fo) ]} .
with the solvate region. Hydrogen atoms that belong to the
water were located on the electron density difference map, and
the H-O bond lengths and the H-O-H bond angles were
constrained to reasonable values. Where possible, suitable
hydrogen acceptors were located using the program CalcOH51 and the H-O vectors positioned to form hydrogenbonding interactions. The coordinates of the hydrogen atoms
belonging to water were allowed to freely refine with the thermal
parameter set to a scaled value of the associated oxygen atom. A
small number of reflections (in some cases) were removed when
(51) Nardelli, M. J. Appl. Crystallogr. 1999, 32, 563–571.
Article
Δ(F2o-F2c )/σ exceeded 10.0. Key data for all structures are given
in Tables 7 and 8. Drawings in Figures 3 through 5 were
produced with the program Diamond 3.1e.52
Cell Line and Culture Conditions. The human A2780 ovarian
cancer cell line was obtained from the European Collection of
Cell Cultures (Salisbury, UK). Cells were grown routinely in
RPMI medium containing glucose, 5% fetal calf serum (FCS),
and antibiotics at 37 °C and 5% CO2.
Cytotoxicity Test (MTT assay). Cytotoxicity was determined
using the MTT assay (MTT = 3-(4,5-dimethyl-2-thiazolyl)-2,5diphenyl-2H-tetrazolium bromide). Cells were seeded in 96-well
plates as monolayers with 100 μL of cell solution (approximately
20 000 cells) per well and preincubated for 24 h in medium
supplemeted with 10% FCS. Compounds were dissolved directly in the culture medium, with the exception of 4b, which was
added as a DMSO solution, and serially diluted to the appropriate concentration (final DMSO concentration = 0.5%).
Then 100 μL of drug solution was added to each well and the
plates were incubated for another 72 h. Subsequently, MTT
(52) Diamond: Crystal and Molecular Structure Visualization, version
3.1e; Crystal Impact: Bonn, 2007.
Organometallics, Vol. 28, No. 17, 2009
5071
(5 mg/mL solution in phosphate-buffered saline) was added to
the cells, and the plates were incubated for a further 2 h. The
culture medium was aspirated, and the purple formazan crystals
formed by the mitochondrial dehydrogenase activity of vital
cells were dissolved in DMSO. The optical density, directly
proportional to the number of surviving cells, was quantified
at 540 nm using a multiwell plate reader, and the fraction of
surviving cells was calculated from the absorbance of untreated
control cells. Evaluation is based on means from two independent experiments, each comprising three microcultures per
concentration level.
Acknowledgment. We thank the Swiss National
Science Foundation and EPFL for financial support.
This research was also supported by a grant from the
European Commission Marie Curie Action (ADP, Project CARCAS, MEIF-CT-2005-025287).
Supporting Information Available: CIF files giving crystallographic data for 1, 2a-d, 4a, and 4c. This material is available
free of charge via the Internet at http://pubs.acs.org.