Insights into the reactivity of gold-dithiocarbamato anticancer agents toward model biomolecules by using multinuclear NMR spectroscopy

Article abstract of DOI:10.1002/chem.201302550

Some gold(III)-dithiocarbamato derivatives of either single amino acids or oligopeptides have shown promise as potential anticancer agents, but their capability to interact with biologically relevant macromolecules is still poorly understood. We investigated the affinity of the representative complex [Au ^IIIBr₂(dtc-Sar-OCH₃)] (dtc: dithiocarbamate; Sar: sarcosine (N-methylglycine)) with selected model molecules for histidine-, methionine-, and cysteine-rich proteins (that is, 1-methylimidazole, dimethylsulfide, and N-acetyl-L-cysteine, respectively). In particular, detailed mono- and multinuclear NMR studies, in combination with multiple ¹³C/¹⁵N enrichments, allowed interactions to be followed over time and indicated somewhat unexpected reaction pathways. Whereas dimethylsulfide proved to be unreactive, a sudden multistep redox reaction occurred in the presence of the other potential sulfur donor, N-acetyl-L-cysteine (confirmed if glutathione was used instead). On the other hand, 1-methylimidazole underwent an unprecedented acid-base reaction with the gold(III) complex, rather than the expected coordination to the metal center by replacing, for instance, a bromide. Our results are discussed herein and compared with the data available in the literature on related complexes; our findings confirm that the peculiar reactivity of gold(III)-dithiocarbamato complexes can lead to novel reaction pathways and, therefore, to new cytotoxic mechanisms in cancer cells. To react or not to react? Model nitrogen- and sulfur-donor compounds can be treated with a representative gold(III)- dithiocarbamato anticancer agent to evaluate potential reactivity of such anticancer agents toward biomolecules (see figure). Detailed NMR studies show that the interaction with physiologically relevant species can lead to unexpected reaction pathways that may contribute to the novel mechanisms of cytotoxicity observed for this class of complexes. Copyright

Full text of DOI:10.1002/chem.201302550

DOI: 10.1002/chem.201302550
Insights into the Reactivity of Gold–Dithiocarbamato Anticancer Agents
toward Model Biomolecules by Using Multinuclear NMR Spectroscopy
Giulia Boscutti,[a] Luciano Marchiꢀ,[b] Luca Ronconi,*[a, c] and Dolores Fregona*[a]
Abstract: Some gold
(III)–dithiocarba-
particular, detailed mono- and multinu-
clear NMR studies, in combination
with multiple 13C/15N enrichments, al-
lowed interactions to be followed over
time and indicated somewhat unex-
pected reaction pathways. Whereas di-
methylsulfide proved to be unreactive,
a sudden multistep redox reaction oc-
curred in the presence of the other po-
tential sulfur donor, N-acetyl-l-cysteine
(confirmed if glutathione was used in-
stead). On the other hand, 1-methyl-
AHCTUNGTRENNUNG
mato derivatives of either single amino
acids or oligopeptides have shown
promise as potential anticancer agents,
but their capability to interact with bio-
logically relevant macromolecules is
still poorly understood. We investigat-
ed the affinity of the representative
complex [AuIIIBr2(dtc-Sar-OCH3)] (dtc:
dithiocarbamate; Sar: sarcosine (N-
methylglycine)) with selected model
molecules for histidine-, methionine-,
and cysteine-rich proteins (that is, 1-
methylimidazole, dimethylsulfide, and
N-acetyl-l-cysteine, respectively). In
ACHTUNGTRENNUNG
complex, rather than the expected co-
ordination to the metal center by re-
placing, for instance, a bromide. Our
results are discussed herein and com-
pared with the data available in the lit-
erature on related complexes; our find-
ings confirm that the peculiar reactivity
of goldACTHNUGTRNEUNG(III)–dithiocarbamato com-
Keywords: antitumor
dithiocarbamates gold
spectroscopy
agents
·
plexes can lead to novel reaction path-
ways and, therefore, to new cytotoxic
mechanisms in cancer cells.
A
·
·
NMR
Introduction
have been used for centuries for the treatment of rheuma-
toid arthritis,[4] their well-known anti-inflammatory and im-
munosuppressing properties have made them promising can-
didates as potential drugs in oncology.
Although cisplatin and its analogues continue to enjoy the
status of the worldꢀs leading anticancer drugs,[1] a major goal
of modern medicinal chemistry and drug design is the devel-
opment of novel metallodrugs with pharmacological profiles
that are different from those of the clinically established
platinum drugs, so as to reduce side effects and enhance the
therapeutic outcomes.[2] Among these metallodrugs, gold
compounds have been carving out a space due to their
strong inhibitory effects on tumor cell growth, which are
generally achieved by exploiting non-cisplatin-like mecha-
In this context, we have been developing some gold
dithiocarbamato derivatives that are designed to combine
the potential cytotoxic properties of the gold(III) ion with
ACHTUNGTRENNUNG(III)–
AHCTUNGTRENNUNG
the likely chemoprotective function of the dithiocarbamato
ligand. In particular, a well-known major limiting factor in
the use of platinum-based therapeutics in the clinic is renal
toxicity.[5] In fact, strong and irreversible binding of platinum
to sulfur-containing biomolecules (including the thiol groups
of renal enzymes) is believed to alter the conformation of
the biomolecules themselves and to lead to substantial
changes in their biological functions, up to inactivation, thus
inducing gastro-, nephro-, and bone marrow toxicity.[6] In
this regard, dithiocarbamates have been successfully used as
chemoprotectants to modulate cisplatin nephrotoxic side ef-
fects without decreasing its antitumor activity,[7] although
the overall nephroprotective benefits were somewhat limit-
ed by the acute toxicity profile, especially at the neuronal
level, exhibited by the free (that is, not-coordinated) dithio-
carbamates.[8] Consequently, metal–dithiocarbamato deriva-
tives are aimed at combining the chemoprotective activity
of the bound dithiocarbamate with the prevention of the in-
trinsic toxicity of the free ligand. Moreover, in the case of
nisms of action.[3] Gold
ACTHNUTRGNEU(GN III) and platinum(II) complexes
share some important chemical features (that is, they are
isoelectronic and isostructural), and since gold compounds
[a] G. Boscutti, Dr. L. Ronconi, Prof. D. Fregona
Department of Chemical Sciences, University of Padova
Via F. Marzolo 1, Padova 35131 (Italy)
E-mail: luca.ronconi@unipd.it
dolores.fregona@unipd.it
[b] Dr. L. Marchiꢁ
Department of Chemistry, University of Parma
Parco Area delle Scienze 17A, Parma 43100 (Italy)
[c] Dr. L. Ronconi
Current address:
School of Chemistry, National University of Ireland Galway
University Road, Galway (Ireland)
E-mail: luca.ronconi@nuigalway.ie
goldACTHNUGRTENUNG(III) complexes, the chelating effect of the dithiocarba-
mato moiety is likely to stabilize the metal atom in the +3
oxidation state and, thus, prevent reduction to gold(I) spe-
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.201302550.
cies or metallic gold (a well-known drawback of goldACHTUNGTRENNUNG(III)
13428
ꢂ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2013, 19, 13428 – 13436

FULL PAPER
derivatives often occurring under the normally reducing
physiological conditions).[9]
Accordingly, we have designed a number of goldACTHNUTRGENUG(N III)–di-
thiocarbamato complexes of the type [AuIIIX2(dtc-Sar-OR)]
(X: Cl, Br; dtc: dithiocarbamate; Sar: sarcosine (N-methyl-
glycine); R: Me, Et, tBu), which were proved to be promis-
ing potential anticancer agents, with excellent cytotoxic ac-
tivity in vitro, subsequently confirmed in vivo,[10] coupled
with negligible acute toxicity and an almost complete lack
of nephrotoxic side effects.[11] However, although insights
into the mechanism of action and biological targets of these
complexes have been recently reported,[12] the lack of actual
mechanistic studies is currently a major issue, particularly
with regard to the affinity of the complexes toward biologi-
cally relevant macromolecules.
Figure 2. 1H NMR spectrum of complex 1 over time, in [D6]acetone
A) soon after dissolution and B) after 24 h, and in [D6]DMSO C) soon
after dissolution and D) after 24 h. DMSO: dimethylsulfoxide.
Proteins are commonly regarded as primary targets for cy-
totoxic goldACHTUNGTRENNUNG(III) compounds because they can provide po-
tential metal-binding sites. These consist mainly of the
donor atoms of the side chains of amino acid residues, in-
cluding the imidazole nitrogen atom in l-histidine and the
thiolate and thioether sulfur atoms in l-cysteine and l-me-
thionine, respectively.[13] Therefore, we investigated the in-
4.81 ppm were recorded in [D6]acetone (Figure 2A) for the
N-methyl and N-methylene groups, respectively. The two
NCH3/NCH2 sets at d=3.58/4.79 and 3.61/4.81 ppm belong
to two different species, as confirmed by the mutual long-
range couplings observed in the [1H,13C] HMBC spectrum
(Figure 3), which indicates the coexistence of two forms of
teraction of the representative goldACTHNUTRGNEUNG(III)–dithiocarbamato
complex [AuIIIBr2(dtc-Sar-OCH3)] (1; Figure 1) with selected
molecules, namely 1-methylimidazole (1-MeIm), dimethyl-
sulfide (DMS), N-acetyl-l-cysteine (l-NAC), and gluta-
thione (GSH, g-Glu-Cys-Gly), which may be regarded as re-
liable models of target biomolecules containing accessible
nitrogen and sulfur donors (that is, histidine-, methionine-,
and cysteine-rich proteins). Somewhat unexpected and un-
precedented results, obtained by means of detailed mono-
and multinuclear NMR studies, are reported and discussed
herein, and possible reaction pathways are proposed.
1
complex 1. In fact, the H signal of the NCH3 group at d=
1
3.58 ppm gives rise to long-range H–13C couplings with the
carbon atoms of the NCH2 (3J) and the CSS (3J) groups (at
d=53.9 and 197.7 ppm, respectively), and the 1H peak of
the NCH2 group at d=4.79 ppm is long-range coupled with
the carbon atoms of the NCH3 (3J), the C=O (2J), and the
CSS (3J) groups (at d=39.9, 166.7, and 197.7 ppm, respec-
tively), all of which belong to the same major species. Anal-
1
ogously, the H signal of the NCH3 group at d=3.61 ppm
gives rise to long-range 1H–13C couplings with the carbon
Results and Discussion
atoms of the NCH2 (3J) and the CSS (3J) groups (at d=54.9
1
and 201.5 ppm, respectively), and the H peak of the NCH2
NMR characterization of complex 1: As previously report-
ed, compound 1 may exist in two forms, the relative abun-
dance of which varies depending on the synthetic condi-
tions.[14] Two proton signals at d=3.58/3.61 and 4.79/
group at d=4.81 ppm is long-range coupled with the carbon
atoms of the NCH3 (3J), the C=O (2J), and the CSS (3J)
groups (at d=40.9, 166.8 and 201.5 ppm, respectively), all of
which belong to the same minor species.
Figure 1. Structures of the goldACHTUNGTRNEUNG
(III)–dithiocarbamato complex [AuIIIBr2(dtc-Sar-OCH3)] (1), the 13C,15N-enriched analogue (13C4,15N-1), the corresponding
gold(I) derivative [AuI(dtc-Sar-OCH3)]2 (2), and the model molecules 1-methylimidazole (1-MeIm), N-acetyl-l-cysteine (l-NAC), glutathione (GSH),
and dimethylsulfide (DMS) studied in this work.
Chem. Eur. J. 2013, 19, 13428 – 13436
ꢂ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
13429

L. Ronconi, D. Fregona et al.
species irreversibly interconverted into the major (mono-
meric) one over time, which led to the almost complete dis-
appearance of the minor species (Figure 2A and B). The
same behavior was also observed in CDCl3 and CD2Cl2
(data not shown), whereas no interconversion occurred in
either CD3OD (data not shown) or [D6]DMSO over 24 h
(Figure 2C and D); these results confirm the crucial role
played by the solvent. Interestingly, after the interconver-
sion reached completeness in [D6]acetone, the solvent was
evaporated and the solid redissolved in [D6]DMSO, and
only the NMR signals of the final major species were detect-
ed (data not shown), which thus confirmed that the process
is irreversible.
Interaction with 1-methylimidazole: Complex 1 (or its
13C,15N-enriched analogue 13C4,15N-1) was treated with two
equivalents of 1-MeIm in [D6]DMSO, and the reaction was
followed by 1H NMR spectro-
Figure 3. ACHTUNGTRENNUNG
[1H,13C] HMBC spectrum of complex 1 in [D6]acetone (dotted
1
arrows indicate long-range H–13C couplings).
scopy over time (Figure S2 in
the Supporting Information).
As shown in Figure 5, both un-
reacted 1 (§) and the gold(I)
*
dimer 2 ( ) were detected as
minor products. Moreover,
peaks at d=3.11, 3.79, and
5.93 ppm (*; integrating for
3:3:1 hydrogen atoms, respec-
tively) have been assigned by
multinuclear NMR techniques
(Figure S3–S5 in the Supporting
Information) to the NCH3,
OCH3, and NCH groups, re-
spectively, of a new major spe-
cies that was identified as the
anion resulting from deprotona-
tion of the N-methylene group
of compound 1. At the same
time, no evidence of coordina-
tion by 1-MeIm to the metal
Figure 4. Details of the 1H NMR spectrum (CD2Cl2 as solvent) and of the IR spectra of A) the native raw com-
plex 1 and B) the major and C) minor forms, as isolated by preparative layer chromatography.
atom was observed, and the
proton signals recorded for the
imidazole moiety (Figure 5C)
are consistent with its N3-proto-
The two forms were isolated by preparative layer chroma-
nated form, 1-MeImH+, which thus acts as the counterion of
1
tography and characterized by FTIR and H NMR spectro-
the goldACTHNGUTER(NNUG III) anionic species resulting from an acid–base re-
scopy (Figure 4). The major species was identified as the ex-
pected goldACHTUNGTRENNUNG(III)–dithiocarbamato derivative 1, depicted in
Figure 1, the structure of which was confirmed by X-ray
analysis (Figure S1 in the Supporting Information). The
action. The presence of an ionic pair was confirmed by con-
ductivity measurements of the sample, for which a molar
conductivity (LM) value of 61.2 WÀ1 cm2 molÀ1 was recorded,
indicating 1:1 electrolytes in DMSO.[15]
minor species is consistent with a gold
two bridging dithiocarbamato ligands are bound to both
gold(III) atoms, the coordination sphere of which is com-
pleted by two trans-bromide moieties, as suggested by the
(III) dimer, in which
This was also confirmed by the reaction with the so-called
Proton Sponge (1,8-bis(dimethylamino)naphthalene, PS). PS
is among the strongest amine bases in water, although its ba-
ACHTUNGTRENNUNG
sicity decreases in DMSO (pKaACTHNUTRGNEUNG
(PSH+)=12.1 (H2O)/7.5
absence of both symmetric S Au S and Br Au Br stretch-
ing vibrations in the far-IR spectrum (Figure 4C).
(DMSO)),[16] but it is a weak nucleophile due to the steric
hindrance provided by the N-methyl groups. Upon treat-
ment of 1 with two equivalents of PS in [D6]DMSO (Fig-
À
À
À
À
1
Intriguingly, when the H NMR spectra of the native com-
1
plex 1 were recorded in [D6]acetone, the minor (dimeric)
ure S6 in the Supporting Information), the H NMR results
13430
www.chemeurj.org
ꢂ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2013, 19, 13428 – 13436

Reactivity of Gold–Dithiocarbamato Anticancer Agents
FULL PAPER
reacted 13C4,15N-1 or l-NAC
were detected (Figure 6). On
the other hand, two distinct
species were identified as the
final products, that is, the oxi-
dized form of l-NAC, N-acetyl-
^
l-cystine (l-NACy; ), and the
13C,15N-enriched
protonated
methylsarcosine (MSHBr, $), as
confirmed by [1H,13C] HMQC
and HMBC (data not shown)
and [1H,15N] HMBC spectrosco-
py (Figure S9 in the Supporting
Information). Remarkably, no
evidence of any gold–dithio-
AHCTUNGTRENNUNG
a
192.8 ppm was recorded in the
13C{1H} NMR spectrum of the
sample, resulting from the pres-
ence of free CS2 in solution
(Figure S10 in the Supporting
Information).
Figure 5. 1H NMR spectra in [D6]DMSO of A) compound 1 (§), B) 1-MeIm, C) compound 1 and 1-MeIm (1:2)
*
after 24 h of mixing (*: new species formed), and D) compound 2 ( ).
Altogether, our results indi-
resembled those obtained with 1-MeIm (although the inter-
action was slower), with the main final reaction products
cate that l-NAC rapidly reacts with 1 and allowed us to
speculate on the possible reaction pathway, which likely in-
volves a multistep redox process in which the l-NAC is oxi-
being the deprotonated gold
ACHTUNGTRNE(NUNG III)–dithiocarbamato deriva-
tive [AuIIIBr (CH3)CHC(O)OCH3)]À and the proto-
2A(SSCN
U
dized to l-NACy and the goldACTHNGUTRENNU(G III) atom undergoes reduc-
nated PS (PSH+). The formation of PSH+ was also con-
firmed by X-ray analysis (Figure S7 in the Supporting Infor-
mation), and the corresponding crystal structure is consis-
tent with that already reported.[17] Remarkably, the same re-
sults were obtained in the
tion to a gold(I) species. Intriguingly, the major reduction
product is not the gold(I) dimer, 2. In fact, apart from
l-NACy, the other species identified are free CS2 and free
protonated methylsarcosine (owing to the release of HBr in
presence of other bases, such as
Et3N (data not shown).
Interaction with N-acetyl-l-cys-
teine and glutathione: We next
investigated the possible reac-
tions in the presence of l-NAC
as a model for thiol-contain-
ing biomolecules. Compound
13C4,15N-1 was treated with two
equivalents of l-NAC in
[D6]DMSO, and the reaction
1
was followed by H NMR spec-
troscopy over time (Figure S8
in the Supporting Information).
Soon after mixing, the initial
orange solution turned immedi-
ately to dark red, which indicat-
ed that an immediate reaction
was taking place. After four
hours, the process reached com-
Figure 6. Comparison between the 1H NMR spectra in [D6]DMSO of A) compound 13C4,15N-1, B) l-NAC,
pleteness and the solution ap-
peared almost colorless. No sig-
C) compound 13C4,15N-1 and l-NAC (1:2) after 4 h of mixing, D) N-acetyl-l-cystine (l-NACy; ), and E) the
^
13C,15N-enriched protonated methylsarcosine (MSHBr; $; obtained by treatment of methylsarcosine hydro-
nals attributable to either un- chloride with an excess of HBr directly in the NMR tube).
Chem. Eur. J. 2013, 19, 13428 – 13436
ꢂ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
13431

L. Ronconi, D. Fregona et al.
solution). Therefore, from a rig-
orous stoichiometric point of
view, the remaining gold(I) spe-
cies is likely to be [AuBr]. Not-
withstanding that the proposed
transient intermediates were
not identified, such a hypothesis
is in agreement with the few
mechanistic data reported in
the literature[18] and is consis-
tent with the complexity of the
NMR spectra recorded soon
after mixing to achieve, eventu-
ally, a relatively simple final
NMR spectrum, the signals of
which have been associated
with well-defined chemical
species.
If glutathione was used in-
stead of l-NAC under the same
experimental conditions, sub-
1
Figure 7. Comparison between the H NMR spectra in [D6]DMSO of A) compound 1 and GSH (1:2) after 4 h
~
;
of mixing, B) the protonated oxidized form of glutathione (GSSGHBr;
obtained by treatment of
*
stantially similar results were glutathione disulfide (GSSG) with HBr (2 equiv) directly in the NMR tube), C) compound 2 ( ), and
obtained. Three 1H NMR de-
D) methylsarcosine (#).
tectable species were identified
after four hours of mixing
(Figure 7): the protonated oxidized form of glutathione
the NMR spectrum in [D6]DMSO.[21] On the contrary, upon
mixing of compound 1 with one equivalent of DMS, no in-
teraction took place (and neither were new species formed)
within 24 h (Figure 8).
~
(GSSGHBr; ; Figure S11 in the Supporting Information),
*
free methylsarcosine (MS; #), and gold(I) dimer 2 ( ). In
contrast to the reaction with l-NAC, protonation occurs for
glutathione disulfide (GSSG) rather than for MS owing to
the higher basicity of the N terminus of GSSG (pKa-
A
Conclusion
of MS (pKaACHTUNGTRENNUNG
(MSH+)=7.44).[20] Moreover, gold(I) dimer 2
turned out to be an additional reduction product, together
with [AuBr], which suggests that the “decomposition by re-
duction” of 1 in the presence of
GoldACTHGNUTERNNU(G III)–dithiocarbamato derivatives have shown promise
as chemotherapeutic agents, and some are now candidates
GSH is not as straightforward
as it is in presence of l-NAC.
Anyway, the interaction with
GSH reached completeness
within four hours and led to the
total reduction of the AuIII spe-
cies to an AuI species.
Interaction with dimethylsul-
fide: Finally, we evaluated the
interaction with DMS as
model of protein methionine
residues. Some gold(III) com-
a
AHCTUNGTRENNUNG
pounds are known to undergo
reduction by oxidizing methio-
nine to methionine sulfoxide.[13]
Accordingly, if such redox reac-
tion occurred, DMS would be
expected to oxidize to DMSO
1
Figure 8. Comparison between the 1H NMR spectra in [D6]DMSO of A) compound 1 (§), B) DMS (%), and
and the corresponding H peak
would appear at d=2.54 ppm in C) compound 1 and DMS (1:1) after 24 h of mixing.
13432
www.chemeurj.org
ꢂ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2013, 19, 13428 – 13436

Reactivity of Gold–Dithiocarbamato Anticancer Agents
FULL PAPER
for entering Phase I clinical trials.[22] In this context, the anti-
cancer activity in vitro of complex [AuIIIBr2(dtc-Sar-OCH3)]
(1) was evaluated in a number of human tumor cell lines
and it showed IC50 values in the low micromolar range and
up to tenfold lower than the reference drug cisplatin.[14,23]
The cytotoxicity levels of 1, as well as the solution proper-
atives are known to be weakly acidic,[27] and their pKa value
decreases upon N,O coordination to transition-metal ions.[28]
Such acidity is expected to further increase in metal–dithio-
carbamato derivatives of amino acids owing to the strong
delocalization of electrons in the NCSS moiety, which indu-
ces a partial positive charge on the nitrogen atom and, thus,
[29]
À
À À
ties, fully resemble those reported for other goldACTHNUTRGNE(NUG III)–dithio-
further weakens the C H bond in the SSCN C H group.
In the actual case, 1-MeIm does not bind to the gold
carbamato analogues,[24] so it may be regarded as a reliable
representative compound to assess the behavior of this class
of anticancer agents toward potential biological targets.
Accordingly, in the present work, we have exploited
mono- and multinuclear NMR spectroscopy techniques to
evaluate the interaction of compound 1 and its 13C4,15N-en-
riched analogue (13C4,15N-1) with some selected models of
biomolecules, and the results are summarized in Figure 9.
ACHTUNGTRENNUNG(III)
center of complex 1 (for example, by displacing a bromide
ligand, as initially expected) but acts as a base (pKaACHTUNGTRENNUNG(1-
MeImH+)=6.15 in DMSO)[30] by deprotonating the N-
methylene group, which leads to the formation of the anion-
ic derivative [AuIIIBr (CH3)CHC(O)OCH3)]À and the
2ACTHNUTRGNENG(U SSCNACHTUTGNRENNUGN
N3-protonated form of 1-methylimidazole (1-MeImH+) as a
counterion (Figure 9A). To the best of our knowledge, this
is the first time that such a deprotonation process has been
reported for gold derivatives of amino acids. Accordingly,
we may assume that protein histidine residues are unlikely
to be potential ligands for this class of metal complexes. For
the sake of completeness, we also checked the stability of
compound 1 under acidic conditions and no changes were
observed over time (data not shown). The microenviron-
ment in tumors is generally acidic,[31] so it is expected that
In order to provide insights into the affinity of goldACTHUNTGRNEUNG(III)–
dithiocarbamato derivatives for physiologically relevant bio-
molecules, we first studied the interaction with 1-MeIm as a
model of protein histidine residues. The amino acid l-histi-
dine is probably the most important metal-binding site in bi-
ological systems, because it is found in the active centers of
many enzymes.[25] With regard to the reactivity toward
goldACHTUNGTRENNUNG(III) complexes, histidine can act as a bidentate ligand
by coordinating the metal atom through the amino and the
N3 imidazole nitrogen atoms. Additionally, it can undergo
redox processes, in which the C2H group of the imidazole
ring is oxidized to a C2=O group, along with the reduction
of the AuIII atom to an Au0 atom.[26] Nevertheless, according
to our results, none of the above-mentioned processes was
observed. On the other hand, a protons of amino acid deriv-
the investigated goldACTHUGNTERN(UNG III)–dithiocarbamato derivatives
would not undergo such an acid–base reaction at the tumor
site.
After the discovery of the antiarthritic properties of
gold(I) thiolates,[4] there has been increasing interest in the
interaction between gold complexes and sulfur-containing
amino acids because it was thought to be responsible for the
Figure 9. Possible pathways for reactions of complex 1 with A) 1-MeIm, B) l-NAC, and C) DMS.
Chem. Eur. J. 2013, 19, 13428 – 13436
ꢂ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
13433

L. Ronconi, D. Fregona et al.
ACTHNUTRGNEUNG
CH3O13C(O)13CH215N(13CH3)H·HCl were used as the sources of 13C and
toxic side effects encountered in chrysotherapy.[32] Gold
ACTHNUTRGNE(NUG III)
15N. The gold(I)–dithiocarbamato complex [AuI(dtc-Sar-OCH3)]2 (2) was
complexes are known to undergo reduction to gold(I) spe-
cies in aqueous solution in the presence of thiol-containing
compounds, such as l-cysteine and glutathione (which con-
sequently oxidize to the corresponding disulfide forms),
with reduction rates that depend on the redox stability in-
duced by the coordinated ligands.[13] In this regard, complex
1 is no exception (Figure 9B) because both l-NAC and
GSH are oxidized to l-NACy and GSSG, respectively, to-
gether with reduction of the AuIII center to an AuI species,
with [AuBr] likely being the major reduced gold(I)-contain-
ing species. Remarkably, it is generally acknowledged that,
in the final resulting gold(I) derivatives, the anionic ligands
are often preserved together with the binding of one
l-NAC/GSH through the cysteine thiolate,[33] whereas there
is no evidence of this for compound 1. Glutathione is among
the most abundant intracellular nonprotein thiols and is in-
volved in various physiological defense functions in cells, in-
cluding detoxification and inactivation of drugs,[34] so such
redox processes might be involved in the mechanism of the
anticancer activity.
prepared as previously described.[24a]
AHCTUNGTRENNUNG
AHCTUNGTRENNUNG
powder; m.p. 1888C (decomp.); elemental analysis: calcd (%) for
C5H8AuBr2NO2S2 (MW=535.03 gmolÀ1): C 11.22, H 1.51, N 2.62, S:
11.99; found: C 11.37, H 1.41, N 2.80, S 11.99; IR (KBr): n˜ =1742 (n˜
(C=
À1
À
À
O)), 1564 (n˜
(nujol): n˜484 (n˜
(BrAuBr)); 1H NMR (300 MHz, [D6]DMSO, 298 K, tetramethylsilane
(TMS)): d=3.40/3.43 (2s, 1J(13C,H)=144 Hz, 2J(15N,H)=4 Hz, 3H;
NCH3), 3.76 (s, 3H; OCH3), 4.77/4.79 ppm (2s, 1J(13C,H)=146 Hz,
2J(13C,H)=12 Hz, 2J(15N,H)=5 Hz, 2H; NCH2); 13C{1H} NMR
(75.48 MHz, [D6]DMSO, 298 K, TMS): d=39.9/40.9 (1J(15N,13C)=6 Hz,
2J(13C,13C)=3 Hz; NCH3), 53.1/53.4 (1J(13C,13C)=62 Hz, 1J(15N,13C)=
(15N,13C)=
U
(C OCH3)), 1016 cm (n˜
E
N
(SAuS)), 252/228 cmÀ1 (n˜as/s
as/sACHTGNUTERNNUNG -
AHCTUNGTRENNUNG
A
ACHTUNGTRENNUNG
AHCTUNGTRENNUNG
A
ACHTUNGTRENNUNG
AHCTUNGTRENNUNG
A
E
ACHTUNGTRENNUNG
7 Hz; NCH2), 53.9 (OCH3), 166.2 (COO), 197.7/201.2 ppm (1J
ACHTUNGTRENNUNG
16 Hz; NCSS); 15N NMR (30.41 MHz, [D6]DMSO, 298 K, NH4Cl): d=
144.8/145.5 ppm (NCSS).
AHCTUNGTRENNUNG
[AuI(dtc-Sar-OCH3)]2 (2): Bis[m-[methyl-N-dithiocarboxy-kS:kS’)-N-
methylglycinato]]digold(I); yield: 73%; dark violet powder; m.p. 1678C
(decomp.); elemental analysis: calcd (%) for C10H16Au2N2O4S4 (MW=
750.44 gmolÀ1): C 16.00, H 2.15, N 3.73, S 17.09; found: C 16.12, H 2.08,
À
N 3.63, S 17.18; IR (KBr): n˜ =1737 (n˜ACTHUNRGTNNEUG(C=O)), 1483 (n˜HCATUNGTREN(NGUN N CSS)), 1211
(n˜
(C OCH3)), 1027 cmÀ1 (n˜
E
(SCS)),
À
361 cmÀ1 (n˜
as/sA
Some goldACHTUNGTRENNUNG(III) compounds are known to undergo reduc-
tion in the presence of methionine residues. The first step
was proved to involve the formation of short-lived gold-
ACHTUNGTRENNUNG(III)–methionine intermediates, followed by the release of
d=3.51 (s, 6H; NCH3), 3.72 (s, 6H; OCH3), 4.82 ppm (s, 4H; NCH2);
13C{1H} NMR (75.48 MHz, [D6]DMSO, 298 K, TMS): d=47.0 (NCH3),
52.2 (OCH3), 59.8 (NCH2), 167.7 (COO), 204.8 ppm (NCSS); 15N NMR
(30.41 MHz, [D6]DMSO, 298 K, NH4Cl): d=146.1 ppm (NCSS).
methionine sulfoxide and the formation of gold(I) deriva-
tives.[35] By contrast, complex 1 proved to be completely
inert toward the model of protein methionine residues,
DMS (Figure 9C). This evidence is in agreement with the
stabilization toward reduction of the gold center in the oxi-
dation state +3 provided by the bidentate dithiocarbamato
ligands,[24a,b] although such stabilization does not seem to be
Elemental analysis: CHNS elemental analyses were carried out on either
a Fisons EA1108 or a Thermo Scientific FLASH 2000 CHNS-O micro-
analyzer.
Conductivity measurements: Conductivity measurements were performed
with an Amel-160 conductometer equipped with a Metrohm AG glass
cell containing two parallel platinum electrodes coated with platinum
black (Kcell =0.13 cmÀ1), on freshly prepared DMSO solutions at (25.0Æ
0.1)8C.
sufficient to avoid reduction of the goldACTHNUGRTENUNG(III) complex by
either l-NAC or GSH.
FTIR spectroscopy: FTIR spectra were recorded in nujol on a Nicolet
Nexus 870 spectrophotometer (1000 scans, resolution of 2 cmÀ1) for the
range 50–600 cmÀ1, and in solid KBr on either a Nicolet 55XC or a
Overall, the results reported herein provide insights into
the reactivity of this class of metal-based anticancer agents
and open up new perspectives into the understanding of
their mechanism of action. These results thus represent a
solid starting point for further mechanistic studies, potential-
ly with a number of biologically relevant macromolecules
under physiologically relevant conditions.
Perkin–Elmer 580B spectrophotometer (32 scans, resolution of 2 cmÀ1
)
for the range 400–4000 cmÀ1. Data processing was carried out by using
OMNIC, Version 5.1 (Nicolet Instrument Corp.).
NMR spectroscopy: All NMR spectra were acquired in the appropriate
deuterated solvent at 298 K on a BrukerAvance DRX300 spectrometer
by using a BBI [1H,X] probe head equipped with z-field gradients. Data
processing was carried out by using MestReNova, Version 6.2 (Mestrelab
Research S.L.).
Typical acquisition parameters for 1D 1H NMR spectra (1H:
300.13 MHz): 16 transients, spectral width of 7.5 kHz, with 32000 data
points and a delay time of 5.0 s. Spectra were processed by using expo-
nential weighting with a resolution of 0.5 Hz and a line-broadening
threshold of 0.1 Hz.
Typical acquisition parameters for 1D 13C{1H} NMR spectra (13C:
75.48 MHz): 8000 transients, spectral width of 18.8 kHz, with 32000 data
Experimental Section
Materials: Methylsarcosine hydrochloride, carbon disulfide, 13C-carbon
disulfide, ethyl-2-hydroxyethylsulfide, N-acetyl-l-cysteine (l-NAC), N-
acetyl-l-cystine (l-NACy), 1,8-bis(dimethylamino)naphthalene (Proton
Sponge, PS), dimethylsulfide (DMS), 1-methylimidazole (1-MeIm), gluta-
thione disulfide (GSSG), [D6]acetone, [D6]DMSO, [D2]dichloromethane,
[D1]chloroform, [D4]methanol (Sigma–Aldrich), potassium tetra-
points and
AHCTUNGTRENNUNG
a delay time of 7.0 s. Sequences were optimized for
1J(13C,1H)=145 Hz, and 1H decoupling was achieved by using the
bromoaurateACHTUNGTRENNUNG
(III) dihydrate (Alfa Aesar), 13C3,15N-methylsarcosine hy-
WALTZ16 pulse sequence. Spectra were processed by using exponential
weighting with a resolution of 2.0 Hz and a line-broadening threshold of
2.8 Hz.
drochloride (Isotec), and glutathione (GSH; Calbiochem) were of re-
agent grade or comparable purity and were used as supplied. All other
reagents and solvents were used as purchased without any further
purification.
Typical acquisition parameters for 2D [1H,13C] HMQC NMR spectra
(1H: 300.13/13C: 75.48 MHz): 512 transients of 32 scans per block, spectral
width of 7.5/18.8 kHz, with 1000/1000 data points and a delay time of
Syntheses: The gold
OCH3)] (1) was synthesized as previously reported.[14] For the corre-
ACHTUNGTRENNUNG
(III)–dithiocarbamato complex [AuIIIBr2(dtc-Sar-
ACTHNUTRGNEUNG
2.0 s. Sequences were optimized for 1J(13C,1H)=145 Hz, and 1H decou-
sponding
13C4,15N-labeled
analogue
(13C4,15N-1),
13CS2
and
pling was achieved by using the GARP pulse sequence. Spectra were
13434
www.chemeurj.org
ꢂ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2013, 19, 13428 – 13436

Reactivity of Gold–Dithiocarbamato Anticancer Agents
FULL PAPER
processed by using cosine-square weighting with a resolution of 1.0/
3.0 Hz and a line-broadening threshold of 0.3/1.0 Hz.
Acknowledgements
Typical acquisition parameters for 2D [1H,13C] HMBC NMR spectra (1H:
300.13/13C: 75.48 MHz): 512 transients of 16 scans per block, spectral
width 7.5/18.8 kHz, with 1000/1000 data points and a delay time of 2.0 s.
This work was supported by the European Union (Marie Curie European
Re-Integration Grant 204828—PEPMIDAS to L.R.) and the Fondazione
Cassa di Risparmio di Padova e Rovigo (PhD fellowship to G.B.). The
authors gratefully acknowledge Mr. Loris Calore for elemental analyses
and colleagues at the Consorzio Interuniversitario di Ricerca in Chimica
dei Metalli nei Sistemi Biologici (CIRCMSB) for helpful discussions.
Sequences were optimized for 1J
G
R
1
no H decoupling. Spectra were processed by using sine-square weighting
with a resolution of 0.3/1.0 Hz and a line-broadening threshold of 0.3/
1.0 Hz.
Typical acquisition parameters for 2D [1H,15C] HSQC NMR spectra (1H:
300.13/15C: 30.41 MHz): 512 transients of 16 scans per block, spectral
width of 7.5/12.1 kHz, with 1000/1000 data points and a delay time of
[1] N. J. Wheate, S. Walker, G. E. Craig, R. Oun, Dalton Trans. 2010, 39,
8113–8127.
[2] S. H. van Rijt, P. J. Sadler, Drug Discovery Today 2009, 14, 1089–
1097.
[3] a) S. Nobili, E. Mini, I. Landini, C. Gabbiani, A. Casini, L. Messori,
Med. Chem. Res. 2010, 3, 550–580; b) S. J. Berners-Price, A. Filipov-
ska, Metallomics 2011, 3, 863–873.
[4] a) E. R. T. Tiekink, Inflammopharmacology 2008, 16, 138–142; b) I.
Ott, Coord. Chem. Rev. 2009, 253, 1670–1681.
[5] a) L. Ronconi, D. Fregona, Dalton Trans. 2009, 10670–10680; b) L.
Ronconi, D. Aldinucci, Q. P. Dou, D. Fregona, Anti-Cancer Agents
Med. Chem. 2010, 10, 283–292; c) E. M. Nagy, L. Ronconi, C.
Nardon, D. Fregona, Mini-Rev. Med. Chem. 2012, 12, 1216–1229.
[6] a) N. Pabla, Z. Dong, Kidney Int. 2008, 73, 994–1007; b) J. D. Stew-
art, H. M. Bolt, Arch. Toxicol. 2012, 86, 1155–1156.
[7] H. Shimada, K. Takahashi, T. Funakoshi, S. Kojima, Biol. Pharm.
Bull. 1993, 16, 368–371.
2.0 s. Sequences were optimized for 1J
was achieved by using the WALTZ16 pulse sequence. Spectra were pro-
cessed by using cosine-square weighting with a resolution of 2.2/3.7 Hz
(15N,1H)=70 Hz, and 1H decoupling
ACHTUNGTRENNUNG
ACHTUNGTRENNUNG
and a line-broadening threshold of 0.7/1.0 Hz.
Typical acquisition parameters for 2D [1H,15N] HMBC NMR spectra (1H:
300.13/15N: 30.41 MHz): 512 transients of 16 scans per block, spectral
width 7.5/12.1 kHz, with 1000/1000 data points and a delay time of 2.0 s.
Sequences were optimized for 1J
no H decoupling. Spectra were processed by using sine-square weighting
with a resolution of 2.2/3.7 Hz and a line-broadening threshold of 0.7/
1.0 Hz.
(15N,1H)=70 Hz/nJ
E
1
1H and 13C chemical shifts were referenced to TMS at 0.00 ppm through
internal referencing to the residual peak of the deuterated solvent em-
ployed. 15N cehmical shifts were referenced to external 15NH4Cl (1.5m in
1m HCl in 90:10 H2O/D2O) at 0.00 ppm.
[8] N. Segovia, G. Crovetto, P. Lardelli, M. Espigares, J. Appl. Toxicol.
2002, 22, 353–357.
X-ray crystallography: Single-crystal diffraction data of complex 1 and
the protonated proton sponge bromide ((PSH)Br) were collected at
293(2) K with a Bruker Smart APEXII diffractometer by using MoKa
(l=0.71073 ꢃ) radiation. The unit cell parameters were obtained by
using 60 w frames of 0.58 width and scanned from 3 different zones of re-
ciprocal lattice. Intensity data were integrated from several series of ex-
posure frames (0.38 width) covering the sphere of reciprocal space.[36] Ab-
sorption corrections were applied by using SADABS software,[37] which
gave minimum/maximum transmission factors of 0.536/1.000 and 0.773/
1.000 for complex 1 and (PSH)Br, respectively. The structures were
solved by direct methods (SIR2004)[38] and refined on F2 with full-matrix
least-squares (SHELXL-97),[39] by using the WinGX software package.[40]
Non-hydrogen atoms were refined anisotropically, and hydrogen atoms
were added at their calculated positions. Graphical material was prepared
with ORTEP-3 software for Windows.[41] A summary of data collection
and structure refinement for complex 1 and (PSH)Br is reported in
Table S1 in the Supporting Information. CCDC 932507 (1) contains the
supplementary crystallographic data for this paper. These data can be ob-
tained free of charge from the Cambridge Crystallographic Data Centre
via www.ccdc.cam.ac.uk/data_request/cif.
[9] U. Koelle, A. Laguna, Inorg. Chim. Acta 1999, 290, 44–50.
[10] a) V. Milacic, D. Chen, L. Ronconi, K. R. Landis-Piwowar, D. Frego-
na, Q. P. Dou, Cancer Res. 2006, 66, 10478–10486; b) L. Cattaruzza,
D. Fregona, M. Mongiat, L. Ronconi, A. Fassina, A. Colombatti, D.
Aldinucci, Int. J. Cancer 2011, 128, 206–215.
[11] C. Marzano, L. Ronconi, F. Chiara, M. C. Giron, I. Faustinelli, P.
Cristofori, A. Trevisan, D. Fregona, Int. J. Cancer 2011, 129, 487–
496.
[12] a) X. Zhang, M. Frezza, V. Milacic, L. Ronconi, Y. Fan, C. Bi, D.
Fregona, Q. P. Dou, J. Cell. Biochem. 2010, 109, 162–172; b) F.
Chiara, A. Gambalunga, M. Sciacovelli, A. Nicolli, L. Ronconi, D.
Fregona, P. Bernardi, A. Rasola, A. Trevisan, Cell Death Dis. 2012,
3, e444.
ˇ ´
[13] B. D. Glisic, U. Rychlewska, M. I. Djuran, Dalton Trans. 2012, 41,
6887–6901.
[14] L. Giovagnini, L. Ronconi, D. Aldinucci, D. Lorenzon, S. Sitran, D.
Fregona, J. Med. Chem. 2005, 48, 1588–1595.
[15] W. J. Geary, Coord. Chem. Rev. 1971, 7, 81–122.
[16] R. L. Benoit, D. Lefebvre, M. Frechette, Can. J. Chem. 1987, 65,
996–1001.
[17] D. Pyz˙alska, R. Pyz˙alski, T. Borowiak, J. Cryst. Spectrosc. Res. 1983,
13, 211–220.
[18] P. M. Yangyuoru, J. W. Webb, C. F. Shaw III, J. Inorg. Biochem.
2008, 102, 584–593.
[19] X. D. Yang, Q. Q. Zhang, R. F. Chen, R. F. Shen, Anal. Sci. 2008, 24,
1005–1012.
[20] S.-S. Choi, J.-E. Ko, Bull. Korean Chem. Soc. 2009, 30, 2542–2548.
[21] G. R. Fulmer, A. J. M. Miller, N. H. Sherden, H. E. Gottlieb, A. Nu-
delman, B. M. Stoltz, J. E. Bercaw, K. I. Goldberg, Organometallics
2010, 29, 2176–2179.
[22] D. Fregona, L. Ronconi, C. Marzano, IT 1347835B1, 2008.
[23] D. Aldinucci, D. Lorenzon, L. Stefani, L. Giovagnini, A. Colombat-
ti, D. Fregona, Anti-Cancer Drugs 2007, 18, 323–332.
[24] a) L. Ronconi, L. Giovagnini, C. Marzano, F. Bettio, R. Graziani, G.
Pilloni, D. Fregona, Inorg. Chem. 2005, 44, 1867–1881; b) L. Ronco-
ni, C. Marzano, P. Zanello, M. Corsini, G. Miolo, C. Maccꢄ, A. Tre-
visan, D. Fregona, J. Med. Chem. 2006, 49, 1648–1657; c) G. Boscut-
ti, L. Feltrin, D. Lorenzon, S. Sitran, D. Aldinucci, L. Ronconi, D.
Fregona, Inorg. Chim. Acta 2012, 393, 304–317.
Chromatography: Separation of the two forms of complex 1 was accom-
plished by preparative layer chromatography on silica gel Merck Kiesel-
gel 60F254 precoated glass plates (2 mm thick) in dichloromethane/ethanol
(97:3). Retention factors (Rf) of 0.9 and 0.3 were measured for the major
and minor species, respectively. Once detached from the plate, each frac-
tion was washed with fresh dichloromethane, filtered, and evaporated to
dryness under reduced pressure. Both fractions were then characterized
by FTIR and NMR spectroscopy, together with X-ray crystallography for
the major form of complex 1 (Figure S1, Table S1, and Table S2 in the
Supporting Information).
NMR sample preparation: As a general procedure, for the NMR studies,
20–25 mm solutions of compound 1 (or the corresponding 13C,15N-en-
riched analogue 13C4,15N-1) in [D6]DMSO (0.5 mL) were used. Unless
otherwise stated, samples were placed directly in the NMR tube with
DMS (1 equiv) or 1-MeIm, l-NAC, or GSH (2 equiv; each previously dis-
solved in [D6]DMSO (0.5 mL)). NMR spectra were then recorded soon
after mixing and over time.
Chem. Eur. J. 2013, 19, 13428 – 13436
ꢂ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
13435

L. Ronconi, D. Fregona et al.
[25] a) R. J. Sundberg, R. B. Martin, Chem. Rev. 1974, 74, 471–517; b) S.
Timꢅri, C. Kallay, K. Osz, I. Sovago, K. Varnagy, Dalton Trans. 2009,
1962–1971.
[26] B. A. Al-Maythalony, A. A. Isab, M. I. M. Wazeer, A. Ibdah, Inorg.
Chim. Acta 2010, 363, 3200–3207.
[34] X. Wang, Z. Guo, Anti-Cancer Agents Med. Chem. 2007, 7, 19–34.
ˇ ´
´
[35] A. V. Vujacic, J. Z. Savic, S. P. Sovilj, K. Mꢆszꢅros Szꢆcsꢆnyi, N. To-
ˇ
´
´
´
dorovic, M. Z. Petkovic, V. M. Vasic, Polyhedron 2009, 28, 593–599.
[36] SMART (Control) and SAINT (Integration) Software for CCD Sys-
tems, Bruker AXS, Inc., Madison, WI, USA 1994.
[27] A. Rios, T. L. Amyes, J. P. Richard, J. Am. Chem. Soc. 2000, 122,
9373–9385.
[37] Area-Detector Absorption Correction, Siemens Industrial Automa-
tion, Inc., Madison, WI, USA 1996.
[28] Y. Ilan, Inorg. Chem. 1987, 26, 2454–2459.
[29] S. Wajda, K. Drabent, Bull. Acad. Pol. Sci. Ser. Sci. Chim. 1977, 25,
963–969.
[38] M. C. Burla, R. Caliandro, M. Camalli, B. Carrozzini, G. L. Cascara-
no, L. De Caro, C. Giacovazzo, G. Polidori, R. Spagna, J. Appl.
Crystallogr. 2005, 38, 381–388.
[30] R. L. Benoit, D. Boulet, L. Seguin, M. Frechette, Can. J. Chem.
1985, 63, 1228–1232.
[39] G. M. Sheldrick, SHELX-97, Programs for Crystal Structure Analy-
sis, Release 97–2, University of Gçttingen, Gçttingen, Germany,
2008.
[40] L. J. Farrugia, J. Appl. Crystallogr. 1999, 32, 837–838.
[41] L. J. Farrugia, J. Appl. Crystallogr. 1997, 30, 565–568.
[31] I. F. Tannock, D. Rotin, Cancer Res. 1989, 49, 4373–4384.
[32] H. A. Bird, Ann. Rheum. Dis. 1990, 49, 331–336.
[33] a) T. Zou, C. T. Lum, S. S.-Y. Chui, C.-M. Che, Angew. Chem. 2013,
125, 3002–3005; Angew. Chem. Int. Ed. 2013, 52, 2930–2933; b) C.
Gabbiani, M. A. Cinellu, L. Maiore, L. Massai, F. Scaletti, L. Mes-
sori, Inorg. Chim. Acta 2012, 393, 115–124.
Received: July 2, 2013
Published online: September 3, 2013
13436
www.chemeurj.org
ꢂ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2013, 19, 13428 – 13436