Modification of Platinum(II) Antitumor Complexes with Sulfur Ligands. 2. Reactivity and Nucleotide Binding Properties of Cationic Complexes of the Types [PtCl(diamine) (L)]NO3 and [{PtCl(diamine)}2(L-L)](NO3)2 (L = Monofunctional Thiourea Derivative; L-L = Bifunctional Thiourea Derivative) in Relation to Their Cytotoxicity

Article abstract of DOI:10.1021/ic970421q

The reactions of [PtCl(en)(tmtu)]NO₃ (1) and [PtCl(dach)(tmtu)]NO₃ (2) (en = 1,2-ethanediamine, dach = racemic trans-1,2-cyclohexanediamine, tmtu = 1,1,3,3-tetramethylthiourea) and [{Pt(en)Cl}₂{μ-C₂H₄(NMeCSNMe ₂)₂S,S′}](NO₃)₂ (3) and [{Pt(en)Cl}₂{μ-C₆H₁₂(NMeCSNMe ₂)₂·S,S′}](NO₃) ₂·0.5EtOH (4) with 5′-GMP and r(GpG) and their chemistry in aqueous solution have been investigated by ¹H and ¹⁹⁵Pt NMR spectroscopy. 1 and 2 only form the monofunctional adducts [Pt(en)5′-GMP-N7)(tmtu)] (I) and [Pt(dach)(5′-GMP-N7)(tmtu)] (II), irrespective of an excess of free nucleotide. ¹⁹⁵Pt NMR chemical shifts of -3003 and -2982 ppm, respectively, confirm a [N₃S] mixed-donor environment of platinum. The bulky tmtu ligand in 1 and 2 decreases the rate of hydrolysis of the Pt-Cl bond and the rate of nucleotide binding compared to analogous reactions for cisplatin and structural analogues. The dinuclear complexes 3 and 4 exhibit an unusual rapid intramolecular disproportionation in solution (t_1/2 = 2.5 and 12 h, respectively) which yields [PtCl₂(en)] and [Pt(en)(L-L]²⁺ (L-L = chelating bifunctional thiourea derivative; δ_pt -3454 with L-L = C₂H₄NMeCSNMe₂)₂). Accordingly, 3 forms the mononuclear adducts [Pt(en)(5′-GMP-N7)₂] (III) and [Pt(en){r(GpG)-N7(1)N7(2)}] (IV). Due to the considerably slower rate of decomposition, 4 gives both the dinuclear adduct [{Pt(en)}₂{μ-C₆H₁₂(NMeCSNMe ₂)₂}{μ-r(GpG)-N7(1),N7{2)}] (V) (70%) and IV (30%). The 5′ sugar residue of r(GpG) in IV exhibits an N-type conformation, as commonly observed in bifunctional adducts that are formed between Pt(II) antitumor complexes and dinucleotides. The absence of this structural feature in V supports the formation of a conformationally less restricted macrochelate. Cytotoxicity data for 1-4 in L1210 leukemia are in accordance with the nucleotide-binding properties of 1 and 2 and the aqueous solution chemistry of the dinuclear compounds 3 and 4. The results indicate that structurally modified thiourea ligands may be interesting for their use as alternative, strongly coordinating carrier groups in platinum(II) antitumor complexes.

Full text of DOI:10.1021/ic970421q

Inorg. Chem. 1998, 37, 717-723
717
Modification of Platinum(II) Antitumor Complexes with Sulfur Ligands. 2. Reactivity and
Nucleotide Binding Properties of Cationic Complexes of the Types [PtCl(diamine)(L)]NO3
and [{PtCl(diamine)}2(L-L)](NO3)2 (L ) Monofunctional Thiourea Derivative; L-L )
Bifunctional Thiourea Derivative) in Relation to Their Cytotoxicity
Ulrich Bierbach, John D. Roberts, and Nicholas Farrell*
Department of Chemistry and Massey Cancer Center, Virginia Commonwealth University,
Richmond, Virginia 23284-2006
ReceiVed April 10, 1997
The reactions of [PtCl(en)(tmtu)]NO3 (1) and [PtCl(dach)(tmtu)]NO3 (2) (en ) 1,2-ethanediamine, dach ) racemic
trans-1,2-cyclohexanediamine, tmtu ) 1,1,3,3-tetramethylthiourea) and [{Pt(en)Cl}2{µ-C2H4(NMeCSNMe2)2-
S,S′}](NO3)2 (3) and [{Pt(en)Cl}2{µ-C6H12(NMeCSNMe2)2-S,S′}](NO3)2‚0.5EtOH (4) with 5′-GMP and r(GpG)
and their chemistry in aqueous solution have been investigated by 1H and 195Pt NMR spectroscopy. 1 and 2 only
form the monofunctional adducts [Pt(en)(5′-GMP-N7)(tmtu)] (I) and [Pt(dach)(5′-GMP-N7)(tmtu)] (II), irrespective
of an excess of free nucleotide. 195Pt NMR chemical shifts of -3003 and -2982 ppm, respectively, confirm a
[N3S] mixed-donor environment of platinum. The bulky tmtu ligand in 1 and 2 decreases the rate of hydrolysis
of the Pt-Cl bond and the rate of nucleotide binding compared to analogous reactions for cisplatin and structural
analogues. The dinuclear complexes 3 and 4 exhibit an unusual rapid intramolecular disproportionation in solution
(t1/2 ) 2.5 and 12 h, respectively) which yields [PtCl2(en)] and [Pt(en)(L-L)]2+ (L-L ) chelating bifunctional
thiourea derivative; δPt -3454 with L-L ) C2H4(NMeCSNMe2)2). Accordingly, 3 forms the mononuclear adducts
[Pt(en)(5′-GMP-N7)2] (III) and [Pt(en){r(GpG)-N7(1),N7(2)}] (IV). Due to the considerably slower rate of
decomposition, 4 gives both the dinuclear adduct [{Pt(en)}2{µ-C6H12(NMeCSNMe2)2}{µ-r(GpG)-N7(1),N7(2)}]
(V) (70%) and IV (30%). The 5′ sugar residue of r(GpG) in IV exhibits an N-type conformation, as commonly
observed in bifunctional adducts that are formed between Pt(II) antitumor complexes and dinucleotides. The
absence of this structural feature in V supports the formation of a conformationally less restricted macrochelate.
Cytotoxicity data for 1-4 in L1210 leukemia are in accordance with the nucleotide-binding properties of 1 and
2 and the aqueous solution chemistry of the dinuclear compounds 3 and 4. The results indicate that structurally
modified thiourea ligands may be interesting for their use as alternative, strongly coordinating carrier groups in
platinum(II) antitumor complexes.
Introduction
carry a sulfoxide ligand instead of the second chloride in the
classical compounds show enhanced cytotoxicity in cisplatin-
resistant cell lines. These findings have been attributed to an
altered metabolism (reactivity and cellular uptake) which may
be related to the hydrolytic behavior of the Pt-S bond and the
stereochemical requirements of the sulfur ligand.4 Model studies
on the DNA-binding properties, however, imply that such
species form cis-DDP-like adducts on DNA.7 It can be
concluded that after loss of chloride and sulfoxide, platinum
will covalently bind to the N7 positions of adjacent guanine
bases (1,2 intrastrand cross-link). The structural impact of this
lesion is a directed bend (∼30°) of the double helix toward the
major groove, as has been demonstrated for a platinated
dodecanucleotide in the solid state.8 In contrast, dinuclear
complexes where the Pt(II) centers are bridged by an aliphatic
diamine linker of variable chain length form a different array
of adducts (1,3 and 1,4 GG interstrand cross-links) that are not
The continued interest in platinum-based antitumor com-
pounds is stimulated by the fact that certain tumors are resistant
to the clinically used drug cisplatin (cis-diamminedichloroplati-
num(II), cis-DDP). The rationale behind the design of new
(“third generation”) drugs is that their novel structural features
would result in an alternative mechanism of action distinct from
that of cis-DDP and its direct analogues.1 Second-generation
drugs such as carboplatin2 (cis-diammine(1,1-cyclobutanedi-
carboxylato)platinum(II)) and [PtCl2(dach)]3 (dach ) trans-1,2-
cyclohexanediamine) form the same type of DNA adducts as
cis-DDP and thus are unlikely to overcome resistance that is
associated with an increased repair of these adducts.
To reduce cisplatin resistance and widen the overall spectrum
of activity, nonclassical platinum(II) antitumor compounds have
been developed. Among these are mononuclear and dinuclear
complexes of general formulas [PtCl(diamine)(sulfoxide-S)]+ 4,5
and [{PtCl(NH3)2}2-µ-{H2N(CH2)nNH2}]2+ 6
Complexes that
.
(4) Farrell, N.; Kiley, D. M.; Schmidt, W.; Hacker, M. P. Inorg. Chem.
1990, 29, 397.
(1) Farrell, N. Cancer InVest. 1993, 11, 578.
(5) Landi, J.; Hacker, M. P.; Farrell, N. Inorg. Chim. Acta 1992, 202, 79.
(6) Farrell, N. Comments Inorg. Chem. 1994, 16, 373.
(7) Lempers, E. L. M.; Bloemink, M. J.; Reedijk, J. Inorg. Chem. 1991,
30, 201.
(2) Blommaert, F. A.; Van Dijk-Knijenburg, H. C. M.; Dijt, F. J.; Den
Engelse, L.; Baan, R. A.; Berends, F.; Fichtinger-Schepman, A. M. J.
Biochemistry 1995, 34, 8474.
(3) Page, J. D.; Husain, I.; Sancar, A.; Chaney, S. G. Biochemistry 1990,
29, 1016.
(8) Takahara, P. M.; Frederick, C. A.; Lippard, S. J. J. Am. Chem. Soc.
1996, 118, 12309.
S0020-1669(97)00421-7 CCC: $15.00 © 1998 American Chemical Society
Published on Web 01/31/1998

718 Inorganic Chemistry, Vol. 37, No. 4, 1998
Bierbach et al.
Instrumentation. 1H NMR spectra were recorded at 300 MHz on
a Varian Gemini-300 NMR instrument equipped with a variable-
temperature unit and software for arrayed kinetic experiments. Except
for nucleotide-binding studies and variable-temperature studies, spectra
were obtained at 294 K. Chemical shifts (δ, ppm) are referenced to
Chart 1
DSS (4,4-dimethyl-4-silapentanesulfonic acid). Proton-decoupled 195
-
Pt NMR spectra were taken on a General Electric QE-300 spectrometer
at 64.54 MHz with a spectral window of 166 000 Hz. A 0.1 M solution
of K2[PtCl4] in D2O served as the external standard. Shifts were
recalculated for [PtCl6]2- standard (δ vs Na2[PtCl6] ) δ vs K2[PtCl4]
- 1631).15
Chloride ion concentrations and pH values were determined at 294
K with an Accumet 925 pH/mV meter equipped with a chloride-
sensitive electrode (Fisher Scientific solid-state ISE, Ag/AgCl reference
electrode) and a Beckman combination electrode, respectively. All pH
values taken of D2O solutions are not corrected for the deuterium isotope
effect and will be stated as pH* throughout this paper. Where
necessary, pH values were adjusted with 0.1 M NaOD and DCl
solutions. Elemental analyses were performed by Robertson Microlit
Laboratories, Madison, NJ.
Reactions in the NMR Tube. The reactions between the platinum
complexes 1-4 and the mono- and dinucleotide (5′-GMP, r(GpG)) were
accessible to mononuclear platinum complexes.6,9,10 These long-
range interactions have been shown to effect an irreversible
conversion of right-handed B-form to left-handed Z-form DNA,
a structural impact that might be relevant for the distinct
biological activity of this type of complex.11
performed in 99.996% D2O at 310 K and were followed by H NMR
1
spectroscopy. All solutions were unbuffered (pH* 6.6-6.9). 1H NMR
spectra were recorded at appropriate time intervals with 64 scans per
time point. The first spectrum was acquired when the sample had
adopted the nominal temperature (ca. 10 min). Arrayed spectra were
recorded using the PAD command available in the Varian software.
Quantitation of the progress of the reactions was achieved by integration
of the nonexchangeable H8 proton(s) of the nucleobase(s) and the H1′
proton(s) of the ribose residue(s). The following reactions were
studied: 1 (10 mM) + 5′-GMP (10 mM, 20 mM); 2 (10 mM) + 5′-
GMP (10 mM, 20 mM); 3 (5 mM) + 5′-GMP (10 mM); 3 (5 mM) +
r(GpG) (5 mM); 4 (5 mM) + r(GpG) (5 mM). Furthermore, the
solution stability of 3 and 4 was monitored at 310 K: 3 (5 mM, 50
mM) in D2O and methanol; 4 (5 mM, 50 mM) in D2O.
Larger amounts of the Pt-nucleotide adducts sufficient for 195Pt
NMR studies (cPt ≈ 50 mM) were synthesized in water according to
the above-mentioned conditions and lyophilized.
Cytotoxicity Studies. The in vitro cytotoxicity studies were
performed using standard assays described previously.16 1 and [PtCl2-
(en)] were dissolved in DMF (HPLC grade) and diluted by serial
dilution in saline solution to a final concentration of 0.5% in DMF.
All other compounds were dissolved in saline solution.
In a search for alternative ligand systems, we are currently
examining thiourea derivatives as carrier groups in Pt(II)
antitumor compounds. Toward this objective, the mononuclear
complexes [PtCl(en)(tmtu)]NO3 (1) and [PtCl(dach)(tmtu)]NO3
(2) (en ) 1,2-ethanediamine, dach ) trans-1,2-cyclohexanedi-
amine, tmtu ) 1,1,3,3-tetramethylthiourea) and the dinuclear
species [{Pt(en)Cl}2{µ-C2H4(NMeCSNMe2)2-S,S′}](NO3)2 (3)
and [{Pt(en)Cl}2{µ-C6H12(NMeCSNMe2)2-S,S′}](NO3)2‚0.5EtOH
(4) have been synthesized and fully characterized (Chart 1).12
The aim of the present study was to investigate the effects of
thiourea coordination on the solution chemistry and nucleobase
chemistry of 1-4. Simple model reactions were performed
utilizing 5′-guanosine monophosphate (5′-GMP) and the di-
nucleotide r(GpG) to monitor the affinity of platinum for
guanine-N7, the preferred binding site for Pt-based drugs.13 The
results of this model study are related to in vitro cytotoxicity
data obtained for 1-4 in cisplatin-sensitive and -resistant L1210
leukemia cells. Reactivity features and nucleotide-binding
modes observed at this model nucleobase level proved to be
instructive for the ongoing design of thiourea-containing
platinum drugs and for their mechanism of action.
Results
General Remarks. The reactions and stability tests reported
in this paper were carried out at 310 K near physiological pH.
For reactions involving r(GpG), a pH drop of about 0.5 pH
unit was observed. No attempts were made to control the pH
with a phosphate buffer to avoid the formation of Pt-phosphate
intermediates.17 Platinum coordination to the N7 positions
(Figure 1) of the guanine bases has been established for the
adducts I-V: 1H NMR spectra of these species show signals
for the H8 base protons shifted to lower field relative to those
of the unplatinated nucleobases. Furthermore, these signals do
not show a downfield shift in spectra taken at low pH. This is
in contrast to the “free” nucleotides, where this effect is observed
Experimental Section
Materials and Procedures. Complexes 1-4 were prepared by the
published procedures.12 cis-[PtCl2(NH3)2] and [PtCl2(en)] were syn-
thesized by following a method described by Dhara.14 The mono-
nucleotide 5′-GMP was employed as its disodium salt (Aldrich). The
dinucleotide r(GpG) was available as its triethylammonium salt‚5.5H2O
(Sigma) which was dissolved in a minimum amount of D2O and
lyophilized. All other reagents were purchased from common vendors
and used as supplied.
due to protonation of unplatinated N7 (pKa ) 2.4).18
A
representative pH titration (pH 2-10) is shown for adduct I in
Figure S1 (Supporting Information). For diplatinated r(GpG)
(in adducts IV and V) no downfield shift of the H8 resonances
(9) Zou, Y.; Van Houten, B.; Farrell, N. Biochemistry 1994, 33, 5404.
(10) Farrell, N.; Appleton, T. G.; Qu, Y.; Roberts, J. D.; Soares Fontes, A.
P.; Skov, K. A.; Zou, Y. Biochemistry 1995, 34, 15480.
(11) Kharatishvili, M.; Qu, Y.; Farrell, N. J. Biol. Inorg. Chem., submitted.
(12) Part 1: Bierbach, U.; Hambley, T. W.; Farrell, N. Inorg. Chem. 1998,
37, 708.
(13) Fichtinger-Schepman, A. M. J.; van der Veer, J. L.; den Hartog, J. H.
J.; Lohman, P. H. M.; Reedijk, J. Biochemistry 1985, 24, 707.
(14) Dhara, S. C. Indian J. Chem. 1970, 8, 193.
(15) Kerrison, S. J. S.; Sadler, P. J. J. Magn. Reson. 1978, 31, 321.
(16) Farrell, N.; Qu, Y.; Hacker, M. P. J. Med. Chem. 1990, 33, 2179.
(17) Appleton, T. G.; Berry, R. D.; Davis, C. A.; Hall, J. R.; Kimlin, H.
A. Inorg. Chem. 1984, 23, 3514.
(18) Dijt, F. J.; Canters, G. W.; den Hartog, J. H. J.; Marcelis, A. T. M.;
Reedijk, J. J. Am. Chem. Soc. 1984, 106, 3644.

Pt(II) Antitumor Complexes with Sulfur Ligands
Inorganic Chemistry, Vol. 37, No. 4, 1998 719
1
Table 1. Selected H and 195Pt NMR Dataa for the Adducts I-V Formed in the Reactions between 1-4 and the Nucleotides 5′-GMP and
r(GpG)
δ(1H)/ppm (3JH1′-H2′′/Hz)
adduct, nucleotide
H8
H1′
δ(195Pt)/ppm
[Pt(en)(5′-GMP-N7)(tmtu)] (I)b
8.77
8.748/8.751
8.71
5.95 (6.0)
5.94 (6.3)
5.96 (4.2)
5.92 (6.1)
5.87 (7.2)/6.04 (0.0)
5.95 (6.1)/5.97 (4.8)
5.81 (5.0)/5.89 (5.1)
-3003
-2982
-2648
[Pt(dach)(5′-GMP-N7)(tmtu)] (II)b
[Pt(en)(5′-GMP-N7)2] (III)c
5′-GMPb
8.18
[Pt(en){r(GpG)-N7(1),N7(2)}] (IV)d
8.19/8.43
8.58/8.63
7.94g/8.01h
n.d.e
n.d.e
[{Pt(en)}2{µ-C6H12(NMeCSNMe2)2-S,S′}{µ-r(GpG)-N7(1),N7(2)}] (V)f
r(GpG)d
a 294 K, D2O solution. b pH* 6.8. c pH* 6.7; see also ref 7. d pH* 6.2. e Insufficient solubility. f pH* 6.0. g Assigned to 3′ guanine; see ref 26.
h Assigned to 5′ guanine, see ref 26.
Figure 2. Progress of the reaction of 5′-GMP (10 mM) with [PtCl-
(en)(tmtu)]NO3 (1) (10 mM) followed by 1H NMR spectroscopy.
Conditions: 310 K, D2O solution, pH* 6.8. Assignments: (O) H8 of
unplatinated 5′-GMP; (b) H8 of platinated 5′-GMP in [Pt(en)(5′-GMP)-
(tmtu)] (I); (9) overlapping H1′ doublets of free and platinated 5′-
GMP.
Scheme 1
Figure 1. Structures of 5′-GMP and r(GpG) giving atom numbering.
The arrows indicate the preferred binding site for Pt-based drugs, the
N7 position of guanine.
is observed on going from low to neutral pH. This is in
agreement with the absence of a deprotonation step of the
phosphodiester group, unlike the situation for the 5′ phosphate
(pKa ) 6.2) in the mononucleotide.18 Selective NMR data for
the final products of the reactions described below are given in
Table 1.
Reactions of [PtCl(en)(tmtu)]NO3 (1) and [PtCl(dach)-
(tmtu)]NO3 (2) with 5′-GMP. The reaction of 1 and 2 with
5′-GMP at equimolar concentrations of platinum complex and
nucleotide results in the displacement of the chloro ligand by
platinum. Furthermore, ion-sensitive measurements indicate that
the nucleobase, giving [Pt(en)(5′-GMP-N7)(tmtu)] (I) and [Pt-
tmtu profoundly affects the hydrolytic behavior of the Pt-Cl
(dach)(5′-GMP-N7)(tmtu)] (II) (Scheme 1). The half-times of
bonds in 1 and 2. At room temperature, 1 mM aqueous
product formation are 9 and 11.5 h, respectively. I and II are
solutions of 1 and 2 slowly release ca. 0.5 mM (50% hydrolysis)
the only adducts formed with no detectable (aqua) intermediates
of chloride within 15 h. This is in contrast to the degree and
appearing during the reaction. The progress of the reaction is
rate of hydrolysis of cisplatin under similar conditions.20 This
depicted for the formation of I in Figure 2. The rates of N7
altered aqueous solution chemistry of 1 and 2 may also
binding for 1 and 2 prove to be considerably slower than those
contribute to their relative kinetic inertness in reactions with
5′-GMP. Aqua cations constitute reactive intermediates in the
found for chloro-am(m)ine complexes of Pt(II) (t1/2 ) 4-5 h)
under analogous conditions.19 This effect may be attributed to
the presence of the bulky tmtu ligand, which should be
unfavorable for a nucleophilic attack of the nucleobase on
(19) Bierbach, U.; Farrell, N. Unreported experiments for the monofunc-
tional complex [PtCl(dien)]Cl (dien ) diethylenetriamine).

720 Inorganic Chemistry, Vol. 37, No. 4, 1998
Bierbach et al.
two-step mechanism of platinum binding to nucleic acid
fragments (rate-determining hydrolysis of the Pt-Cl bond
followed by substitution of the aqua ligand by the nitrogen
base).21
Despite the bulkiness of the peralkylated thiourea derivative
in the position cis to the nucleobase in I and II, free rotation
about the Pt-N7 bonds and bonds of the thiourea moieties is
1
observed. This results in sharp H NMR signals for both the
H8 singlets and the H1′ doublets of the sugar residues (see Table
1) and the methyl protons of tmtu at 3.10 ppm. Platinum
coordination to N7 of guanine in mononucleotides is known to
cause a change in the sugar ring conformation from S-type (C2′-
endo, C3′-exo) to N-type (C3′-endo, C2′-exo).22 In I and II,
this transition is not observed, as can be deduced from 3JH1′-H2′
coupling constants (Table 1). A plausible reason for this effect
cannot be given, but steric effects of the thiourea ligand may
play a role. Two H8 signals for II, separated by only 0.003
ppm, are consistent with the formation of a 1:1 mixture of two
diastereomers, produced by the chirality of the R-ribose and the
presence of the R,R and S,S forms of the dach ligand. 195Pt
chemical shifts for I and II around -3000 ppm (vs [PtCl6]2-
)
are in agreement with a [N3Sthiourea] environment of platinum-
(II).23
No reaction is observed when the monofunctional platinum-
nucleotide adducts I and II are incubated with a second
equivalent of 5′-GMP at 310 K for 2 weeks. Similarly, in
experiments that employ a 1:2 molar ratio of platinum and
nucleotide, half of the applied 5′-GMP remains unreacted in
solution. Species I and II do not bind a second nucleobase.
Neither a displacement of the tmtu ligand nor a ring opening
of the diamine chelate (labilization of the Pt-N bond trans to
sulfur) is observed during the long-term incubations at physi-
ological pH.
Solution Behavior of the Dinuclear Complexes 3 and 4.
Stability tests on aqueous solutions of 3 initially led to the
conclusion that this species shows an increased rate of hydrolysis
compared to the mononuclear species 1 and 2 (ion-selective
measurements). 1H NMR spectra taken of 3 in D2O at 310 K
at appropriate time intervals (Figure 3) show the disappearance
of signals at 2.71 ppm (en) and at 3.16, 3.31, and 4.18 ppm
(bridging thiourea ligand) while new signals appear in the 2.6-
2.8 ppm region (en) and at 3.25 and 3.37 ppm. An almost
unrecognizably broadened signal is observed at 3.84 ppm that
slightly sharpens at temperatures above 330 K. From aqueous
solutions with an initial concentration in 3 higher than 50 mM,
bright-yellow needles precipitate which are identified as
[PtCl2(en)] (analyses, IR). The remaining solution contains a
species that gives a 195Pt chemical shift of -3454 ppm,
indicative of a [N2S2] environment of platinum(II).23 According
to these findings, 3 slowly disproportionates into the mono-
nuclear species [PtCl2(en)] (which in fact rapidly hydrolyzes)
and [Pt(en){C2H4(NMeCSNMe2)2-S,S′}]2+, the half-time of this
process being 2.5 h (see caption of Figure 3 for further
explanations). Analogous products are obtained for the di-
nuclear complex 4, but in this case the reaction proceeds much
slower (t1/2 ≈ 12 h).
Figure 3. Disproportionation of [{PtCl(en)}2{µ-C2H4(NMeCSNMe2)2-
S,S′}](NO3)2 (3) followed by 1H NMR spectroscopy. Conditions: 310
K, c ) 5 mM, D2O solution. Assignments: (O) thiourea and en protons
of 3; (b) signals of the decomposition product, [Pt(en){C2H4-
(NMeCSNMe2)2-S,S}](NO3, Cl); (×) signals of the en ligand in [PtCl2-
(en)] and its hydrolysis products. Signal broadenings observed for the
bis(thiourea) chelate (b) are attributed to the slow dynamics of
interconversion of different chelate ring conformations. The arrow
indicates the broad signal assigned to the central ethylene protons of
the thiourea ligand, which are affected most.
Scheme 2. Proposed Mechanism of Decomposition for 3
and 4
Considering the experimental data, we propose a mechanism
for the decomposition of 3 and 4 (Scheme 2) that involves an
intramolecular nucleophilic attack of coordinated thiourea sulfur.
A transition state with the two platinum centers bridged by a
sulfur atom is suggested, which finally rearranges into the
mononuclear complexes. This mechanism is supported by the
following observations: (i) the fact that mononuclear 1 and 2
do not disproportionate at the same concentrations in platinum
and 4 reacts slower than 3 (separation of the Pt centers by an
extended linker) points to an intramolecular process; (ii) the
(20) (a) Greene, R. F.; Chatterji, D. C.; Hiranaka, P. K.; Gallelli, J. F. Am.
J. Hosp. Pharm. 1979, 36, 38. (b) Miller, S. E.; House, D. A. Inorg.
Chim. Acta 1989, 161, 131. (c) Miller, S. E.; House, D. A. Inorg.
Chim. Acta 1989, 166, 189.
(21) Marcelis, A. T. M.; Van Kralingen, C. G.; Reedijk, J. J. Inorg.
Biochem. 1980, 13, 213.
(22) Okamoto, K.; Behnam, V.; Viet, M. T. P.; Polissiou, M.; Gauthier,
J.-Y.; Hanessian, S.; Theophanides, T. Inorg. Chim. Acta 1986, 123,
3.
(23) Appleton, T. G.; Hall, J. R.; Ralph, S. F. Inorg. Chem. 1985, 24, 4685.

Pt(II) Antitumor Complexes with Sulfur Ligands
Inorganic Chemistry, Vol. 37, No. 4, 1998 721
Scheme 3
reaction does not require hydrolysis of the Pt-Cl bond (the same
reactivity is observed in nonaqueous solutions), which should
favor a nondissociative, concerted mechanism and a direct attack
of coordinated sulfur utilizing its vacant lone pair (Scheme 2).
Bridging thiourea sulfur (here in a putative five-coordinate
transition state) is a well-documented binding mode in poly-
nuclear transition metal complexes.24
Reaction of [{Pt(en)Cl}2{µ-C2H4(NMeCSNMe2)2-S,S′}]-
(NO3)2 (3) with 5′-GMP and r(GpG). When 3 is incubated
with 2 equiv of 5′-GMP, the H8 signal of the free nucleotide
decreases in intensity (t1/2 ) 5.5 h) and a single downfield-
shifted H8 peak appears at 8.71 ppm. The only adduct formed
proves to be [Pt(en)(5′-GMP-N7)2] (III), in accordance with
1
published H NMR data7 and a 195Pt shift of -2648 ppm25
(Table 1). H8 resonances at 8.62 and 8.79 ppm that transiently
appear during the early phase of the reaction are assigned to
the intermediates21 [PtCl(en)(5′-GMP-N7)] and [Pt(D2O)(en)-
(5′-GMP-N7)].
The analogous reaction of 3 with r(GpG) solely yields the
mononuclear adduct [Pt(en){r(GpG)-N7(1),N7(2)}] (IV) (Scheme
3), which is in agreement with the results obtained for 5′-GMP.
Figure 4 depicts the progress (t1/2 ) 2 h) of the reaction as
monitored by characteristic changes in the H8 and H1′ proton
resonances. 1H NMR data for IV such as chemical shift
differences between the H8 protons of the 5′ and the 3′ guanine
3
base and values for JH1′-H2′ (Table 1) appear to be almost
identical with those reported for the analogous N7,N7-chelate
formed by cis-[PtCl2(NH3)2].26 The absence of the latter vicinal
coupling is indicative of a switch in the sugar conformation of
the 5′ ribose residue from S-type to N-type (3J ) 0 Hz; see
Figure 4b) upon platination. This is a well-established structural
feature for bifunctional adducts that are formed between
mononuclear platinum complexes and adjacent nucleobases in
dinucleotides.27
Figure 4. Progress of the reaction of r(GpG) (5 mM) with [{PtCl-
(en)}2{µ-C2H4(NMeCSNMe2)2-S,S′}](NO3)2 (3) (5 mM) followed by
1H NMR spectroscopy: (a) H8 proton region; (b) H1′ proton region.
Conditions: 310 K, D2O solution, pH* 6.8-6.2. Assignments: (O)
free r(GpG); (b) platinated r(GpG) in adduct IV. Note the absence of
the vicinal (3JH1′-H2′) coupling for the 5′ sugar H1′ proton, indicating
a fully N-type ribose conformation.
Reaction of [{Pt(en)Cl}2{µ-C6H12(NMeCSNMe2)2-S,S′}]-
(NO3)2‚0.5EtOH (4) with r(GpG). The reaction of r(GpG)
with 4 proceeds significantly slower (t1/2 ≈ 5 h for the
disappearance of free nucleotide) than with 3. The H8 region
1
of the H NMR spectrum taken after 48 h (Figure 5a) shows
four signals downfield of those of free r(GpG). The two least
downfield-shifted H8 signals and the signals detected for the
H1′ sugar protons at 5.87 and 6.04 ppm (Figure 5b) confirm
the formation of [Pt(en){r(GpG)-N7(1),N7(2)}] (IV) (Table 1).
In addition to the mononuclear adduct IV (30%), a new species
has formed (70%), as evidenced by H8 signals at 8.58 and 8.63
ppm and two overlapping H1′ doublets at 5.95 and 5.97 ppm.
We propose the formation of the macrochelate [{Pt(en)}2{µ-
(24) Bierbach, U.; Saak, W.; Haase, D.; Pohl, S. Z. Naturforsch. 1990,
45B, 45 and references therein.
(25) Miller, S. K.; Marzilli, L. G. Inorg. Chem. 1985, 24, 2421.
(26) Girault, J.-P.; Chottard, G.; Lallemand, J.-Y.; Chottard, J.-C. Bio-
chemistry 1982, 21, 7, 1352.
(27) Farrell, N. In Transition Metal Complexes as Drugs and Chemothera-
peutic Agents; Ugo, R., James, B. R., Eds.; Kluwer: Dordrecht, The
Netherlands, 1989; Chapter 4 and literature cited therein.
C6H12(NMeCSNMe2)2-S,S′}{µ-r(GpG)-N7(1),N7(2)}]
(V)
(Scheme 3), in agreement with the integral intensity ratio
observed for H8, H1′ and the en protons at 2.82 ppm. Unlike

722 Inorganic Chemistry, Vol. 37, No. 4, 1998
Bierbach et al.
Table 2. In Vitro Cytotoxicity of 1-4 in L1210 Leukemia Cellsa
ID50/µMb
complex
L1210/0 L1210/DDP
[PtCl(en)(tmtu)]NO3 (1)
[PtCl(dach)(tmtu)]NO3 (2)
cis-[PtCl2(NH3)2] (cisplatin)
>50
>50
0.19
>50
>50
11.63 (61)c
39.58 (9.5)
[{PtCl(en)}2{C2H4(NMeCSNMe2)2}](NO3)2 4.15
(3)
[{PtCl(en)}2{C6H12(NMeCSNMe2)2}](NO3)2 6.93
(4)
56.63 (8.2)
[PtCl2(en)]
1.25
>20 (>16)
a Complexes in saline solution except for 1 and [PtCl2(en)], which
were in 0.5% DMF. b Drug concentration required to inhibit cell growth
by 50%. c Resistance factor, defined as ID50(resistant)/ID50(sensitive),
is given in parentheses.
modified with dinuclear platinum complexes.28,29 The stepwise
reaction of 4 with r(GpG) should involve intermediates with
one platinum bound to either the 3′ or the 5′ guanine base,
followed by closure to the macrochelate (and polymer formation;
see caption of Figure 5). Interestingly, the H8 signal of 5′
guanine in free r(GpG) decreases in intensity more rapidly
(Figure 5a) than that of the 3′ base. This observation and the
fact that only one major intermediate H8 signal is observed after
2 h indicate that 4 preferentially binds to the 5′ N7 position in
the first reaction step. The final bifunctional adduct is stable
in solution under the applied conditions. This proves that the
mononuclear chelate IV is not a decomposition product of the
macrochelate V but is formed by the direct reaction with [PtCl2-
(en)].
Cytotoxicity Studies. The cytotoxicity of complexes 1-4
was studied both in cisplatin-sensitive and in cisplatin-resistant
murine L1210 leukemia cells (Table 2). Solutions of the
dinuclear complexes 3 and 4 were freshly prepared and
incubated immediately after dilution to take the instability of
the drugs into account. Two experiments were performed. In
the first series of incubations, the mononuclear complexes 1
and 2 were studied and compared with cisplatin. In the second
series, the dinuclear complexes 3 and 4 were studied together
with [PtCl2(en)], their potential decomposition product. Pre-
liminary results are given in Table 2. 1 and 2 show no drug
response at drug concentrations <50 mM in both cisplatin-
sensitive (L1210/0) and cisplatin-resistant (L1210/DDP) cells
and thus have to be considered inactive. The dinuclear
complexes 3 and 4 are fairly active in the sensitive cell line but
show a marked decrease in cell growth inhibition in resistant
cells. 3 appears to be slightly more active than 4 with the cross-
resistances to cisplatin being similar for both complexes. [PtCl2-
(en)] shows slightly greater efficacy than 3 and 4 but proves to
be less active than cisplatin and does not overcome cisplatin
resistance in L1210/DDP.
Figure 5. Progress of the reaction of r(GpG) (5 mM) with [{PtCl-
(en)}2{µ-C6H12(NMeCSNMe2)2-S,S′}](NO3)2‚0.5EtOH (4) (5 mM) fol-
lowed by 1H NMR spectroscopy: (a) H8 proton region; (b) H1′ proton
region. Conditions: 310 K, D2O solution, pH* 6.6-6.0. Assignments:
(O) free r(GpG); (b) platinated r(GpG) in the mononuclear adduct IV;
([) platinated r(GpG) in the macrochelate V; (×) monoplatinated
intermediate; see text. The broadening of the peaks is caused by a small
amount of precipitate (of unknown nature) which forms immediately
after combining the reactants in the NMR tube but slowly redissolves
as the reaction proceeds. The formation of polymeric species of the
type -[PtsPt-N7-r(GpG)-N7-]n is not observed. The broad bases
of the H8 signals for V result from unresolved 195Pt-1H couplings.
Discussion
The nucleotide-binding studies on [PtCl(en)(tmtu)]NO3 (1)
and [PtCl(dach)(tmtu)]NO3 (2) unequivocally show that tmtu
does not act as a leaving group in reactions with guanine, which
implies that 1 and 2 will form monofunctional adducts on
double-stranded DNA. This observation seems plausible since
the reverse of this reaction has been observed: (i) thiourea (SC-
(NH2)2, tu) and its partially N-alkylated derivatives have been
shown to slowly displace guanosine (Guo) and 5′-GMP from
(28) Qu, Y.; Bloemink, M. J.; Reedijk, J.; Hambley, T. W.; Farrell, N. J.
Am. Chem. Soc. 1996, 118, 9307.
(29) Mellish, K. J.; Qu, Y.; Scarsdale, N.; Farrell, N. Nucl. Acids Res. 1997,
25, 1265.
3
the situation for IV, the JH1′-H2′ coupling constants (Table 1)
confirm the absence of a SfN conformational change of the 5′
ribose residue as commonly observed in (deoxy)oligonucleotides

Pt(II) Antitumor Complexes with Sulfur Ligands
Inorganic Chemistry, Vol. 37, No. 4, 1998 723
the bifunctional adducts cis-[Pt(NH3)2(Guo)2]2+ (k ≈ 10-5 s-1 30
and cis-[Pt(NH3)2(5′-GMP)2];31 (ii) thiourea is used as a
mechanistic probe which traps monofunctional Pt-DNA adducts
and as a reagent that effectively removes Pt from DNA.32 At
least in the case of cis-DDP related drugs, the critical lesion
that leads to cytotoxicity is thought to be the 1,2 intrastrand
cross-link of guanine bases. We therefore suggest that the
inability of 1 and 2 to form this adduct could be the reason for
the observed inactivity of these complexes in vitro. Mono-
functional Pt-DNA adducts such as those formed by the
inactive complex [PtCl(dien)]+ (dien ) diethylenetriamine) are
known to be ineffective in blocking DNA replication.33
The ligand properties of tmtu clearly distinguish 1 and 2 from
analogous cationic complexes [PtCl(diamine)(SORR′)]+ which
carry a (chiral) S-bound sulfoxide ligand. The substitution
lability of the latter sulfur ligands is undoubtedly a prerequisite
for the biological activity of these species.4,5,7 At this point, a
comparison of ligand properties of different sulfur nucleophiles
that are relevant to platinum antitumor chemistry seems
worthwhile. It appears that in these systems ligands with
π-acceptor properties34,35 (thioether, sulfoxide) are substituted
by N7 of guanine to give the thermodynamically more stable
Pt-nucleobase adduct. The distinct reactivity of the Pt-Sthioether
bond in Pt-methionine adducts in competition reactions with
guanine-N7 and the possible biological implications should be
noted.36,37 On the other hand, for sulfur ligands that lack
acceptor properties but are strong (σ and π) donors, such as
thiourea12,38 and thiolate,39 obviously the reverse reactivity is
observed.
The rationale behind the development of the dinuclear
complexes 3 and 4 was that these would act as cross-linking
agents with the two platinum subsites binding monofunctionally
to DNA. This is a well-established binding mode for analogous
complexes of the formula [{PtCl(NH3)2}2(H2N-R-NH2)]2+ (R
) linear aliphatic chain), which allows the targeting of larger
DNA sequences for bifunctional adduct formation.6 Accord-
ingly, in 3 and 4 the bridging thiourea derivatives should act as
tightly bound, nonreplaceable spacers that determine the “bite”
on DNA. Initial nucleotide-binding studies on 3 and 4 were
performed to monitor the ability of these complexes to form
stable dinuclear units containing the motif [N7-Pt-{µ-bis-
(thiourea)-S,S′}-Pt-N7]. Although reactions with r(GpG)
)
should also yield information about the geometric requirements
for the specific 1,2 intrastrand cross-link for these complexes,
it was not the intention of this study to favor or disfavor this
adduct over alternative long-range adducts.
In reactions between 3 and the nucleotides, both 5′-GMP and
r(GpG) exclusively react with [PtCl2(en)], the rapidly formed
and most reactive species in the reaction media. The short half-
life of 3 under the conditions of the nucleotide-binding
experiments apparently controls the nature of the final adducts.
For this reason, the formation of a “macrochelate” with dinuclear
3 coordinating to the dinucleotide seems unfavorable, irrespec-
tive of the stereochemical feasibility of this adduct. In vitro
cytotoxicity data for 3 are in agreement with [PtCl2(en)] being
the only biologically active species. The results may also
explain the observed cross-resistance of 3 to cisplatin, based
on a similar mechanism of action and an identical array of DNA
adducts for [PtCl2(en)]. Despite the increased stability of the
analogous dinuclear complex 4 and the ability to form a stable
macrochelate (at least in our model studies), an overall similar
activity is found in L1210 leukemia. The cytotoxicity data for
4 imply that in this case mononuclear [PtCl2(en)] also is the
only “surviving” species that reaches the target DNA but
basically do not rule out the possibility of bifunctional adducts
of (intact) 4 with DNA in vitro. The frequency of such cross-
links may be low and/or these adducts may not significantly
contribute to the biological activity of 4.
Conclusions and Perspectives
The specific incorporation of thiourea derivatives as S-donor
ligands into platinum antitumor complexes drastically changes
the reactivity of the Pt(II) center toward the biologically relevant
target guanine-N7. Tmtu in [PtCl(en)(tmtu)]NO3 (1) and [PtCl-
(dach)(tmtu)]NO3 (2) behaves as a typical nonleaving group due
to the thermodynamic stability of the Pt-S bond. Although
undesired in the case of 1 and 2, this structural motif proves to
be extremely valuable for the design of the dinuclear species 3
and 4, where platinum has the same ligand environment.
However the inherent reactivity of these species appears to be
unfavorable for their use as anticancer drugs. Our ongoing
research will focus on how the mononuclear compounds can
be activated despite their monofunctional covalent binding mode
and on structural modifications of the dinuclear compounds to
increase their stability. With the use of bis(thiourea) ligands
that are conformationally more rigid than those in 3 and 4, it
should be possible to avoid chelation of a single Pt center and
consequently intramolecular disproportionation.
(30) Beaty, J. A.; Jones, M. M. Inorg. Chem. 1992, 31, 2547.
(31) Urata, H.; Akagi, M. Biochem. Biophys. Res. Commun. 1989, 161,
819.
(32) Farrell, N. In Transition Metal Complexes as Drugs and Chemothera-
peutic Agents; Ugo, R., James, B. R., Eds.; Kluwer: Dordrecht, The
Netherlands, 1989; Chapter 2 and literature cited therein.
(33) Bruhn, S. L.; Toney, J. H.; Lippard, S. J. In Progress in Inorganic
Chemistry: Bioinorganic Chemistry; Lippard, S. J., Ed.; Wiley: New
York, 1990; pp 495-505.
(34) Murray, S. G.; Hartley, F. R. Chem. ReV. 1981, 81, 365.
(35) Davies, J. A. AdV. Inorg. Chem. Radiochem. 1981, 24, 115.
(36) Barnham, K. J.; Djuran, M. I.; Murdoch, P. d. S.; Sadler, P. J. J. Chem.
Soc., Chem. Commun. 1994, 721.
(37) Barnham, K. J.; Djuran, M. I.; Murdoch, P. d. S.; Ranford, J. D.; Sadler,
P. J. Inorg. Chem. 1996, 35, 1065.
(38) Bierbach, U.; Barklage, W.; Saak, W.; Pohl, S. Z. Naturforsch. 1992,
47B, 1593.
Acknowledgment. This work was supported by a research
fellowship (to U.B.) from the Deutsche Forschungsgemeinschaft
(DFG, Bonn, Germany) and by the American Cancer Society
(Grant No. DHP-2E). We thank S. Kelemen and J. Peroutka
for their technical assistance in obtaining cytotoxicity data.
Thanks are also expressed to W. C. Heraeus GmbH (Hanau,
Germany) for a generous loan of K2[PtCl4].
Supporting Information Available: A plot of chemical shift of
H8 vs pH* for [Pt(en)(5′-GMP-N7)(tmtu)] (adduct I) (1 page).
Ordering information is given on any current masthead page.
(39) Ashby, M. T. Comments Inorg. Chem. 1990, 10, 297.
IC970421Q