Reactivity of auranofin with selenols and thiols - Implications for the anticancer activity of gold(I) compounds

Article abstract of DOI:10.1002/ejic.201300058

The enzyme thioredoxin reductase (TrxR) is attracting much interest as a potential target for cancer therapy. The presence of a selenium atom in the catalytic site makes it sensitive to inhibition by electrophilic molecules, including the Au^I complex auranofin [2,3,4,6-tetra-O-acetyl-1-thio- β-D-glucopyranosato-S-(triethylphosphane)gold]. The reactions between auranofin and models of thiol and selenol nucleophiles present in TrxR (PhSH and PhSeH) have been investigated in chloroform and methanol by means of ¹H, ³¹P, and ⁷⁷Se NMR spectroscopy. In chloroform, auranofin undergoes ligand substitution of the tetraacetylthioglucose moiety by a PhS or PhSe group. The reaction is reversible in both cases, but it is characterized by widely different equilibrium constants (ca. 1 for S and at least 10³ for Se). In polar solvents, such as methanol, the reaction is more complex, and the phosphane moiety also undergoes ligand exchange. Some features have been clarified through the investigation of Et₃PAuCl. The elementary processes involved have been characterized by DFT calculations. Copyright

Full text of DOI:10.1002/ejic.201300058

FULL PAPER
DOI:10.1002/ejic.201300058
Reactivity of Auranofin with Selenols and Thiols –
Implications for the Anticancer Activity of Gold(I)
Compounds
Francesca Di Sarra,[a] Barbara Fresch,[a] Riccardo Bini,[a]
Giacomo Saielli,[b] and Alessandro Bagno*[a]
Keywords: Gold / S ligands / Enzymes / Inhibitors / Medicinal chemistry / Antitumor agents / Auranofin
The enzyme thioredoxin reductase (TrxR) is attracting much
interest as a potential target for cancer therapy. The presence
of a selenium atom in the catalytic site makes it sensitive to
inhibition by electrophilic molecules, including the AuI com-
stitution of the tetraacetylthioglucose moiety by a PhS or
PhSe group. The reaction is reversible in both cases, but it is
characterized by widely different equilibrium constants (ca.
1 for S and at least 103 for Se). In polar solvents, such as
methanol, the reaction is more complex, and the phosphane
moiety also undergoes ligand exchange. Some features have
been clarified through the investigation of Et3PAuCl. The el-
ementary processes involved have been characterized by
DFT calculations.
plex auranofin [2,3,4,6-tetra-O-acetyl-1-thio-β-D-glucopyr-
anosato-S-(triethylphosphane)gold]. The reactions between
auranofin and models of thiol and selenol nucleophiles pres-
ent in TrxR (PhSH and PhSeH) have been investigated in
chloroform and methanol by means of 1H, 31P, and 77Se NMR
spectroscopy. In chloroform, auranofin undergoes ligand sub-
deed, several emerging anticancer therapies have identified
TrxR as a target for drug development,[14] as altered activi-
ties of the Trx system proteins have been observed in several
human diseases.[15]
Many therapeutic compounds have been identified as
TrxR inhibitors, including gold(I) compounds that have
also shown early promise as anticancer drugs.[16–20] The best
known is auranofin [AF, 2,3,4,6-tetra-O-acetyl-1-thio-β-d-
gluco-pyranosato-S-(triethylphosphane)gold, Figure 1], a li-
pophilic gold compound[21] with anti-inflammatory and im-
munosuppressive properties,[1,22] which is used clinically in
the treatment of rheumatoid arthritis.[2,9,23–27]
Introduction
There is an ongoing interest in the identification of new
metal-based drugs for anticancer therapy that have different
mechanisms of action from that exhibited by PtII com-
pounds (formation of cytotoxic lesions with DNA with re-
lated toxic side effects).[1–4] An avenue of investigation that
is attracting much attention is provided by the thioredoxin/
thioredoxin reductase (Trx/TrxR) system.
The Trx/TrxR system is present in all living cells and is
an essential constituent of the cell proliferation machinery
in mammalians.[5–8] It is also an indicator of disorders such
as rheumatoid arthritis, Sjoegren’s syndrome, AIDS, and
certain types of cancer.[9–12]
All mammalian TrxR isoenzymes are selenoproteins with
a redox-active selenocysteine (Sec) residue in their active
sites that catalyzes the reduced nicotinamide adenine dinu-
cleotide phosphate (NADPH) dependent reduction of their
native Trx substrates.[3,13] The presence of selenium in an
enzyme with such critical functions offers unique opportu-
nities for the development of new drugs, as one can exploit
Figure 1. Structure of auranofin.
Auranofin is probably the most effective inhibitor of
the affinity of selenium for electrophilic heavy elements. In- mammalian TrxR found to date.[7,28,29] It inhibits mito-
chondrial thioredoxin reductase (TrxR) and induces mito-
chondrial permeability,[30,31] which results in the release of
[a] Dipartimento di Scienze Chimiche, Università di Padova,
cytochrome c from mitochondria into cytoplasm[32,33] with
Via Marzolo, 35131 Padova, Italy
E-mail: alessandro.bagno@unipd.it
consequent apoptotic cell death.[34] Thus, AF might be
utilized as an anticancer agent to induce apoptotic cell
death.[35,36]
Homepage: http://www.chimica.unipd.it
[b] CNR, Istituto per la Tecnologia delle Membrane, Unità di
Padova,
Via Marzolo, 35131 Padova, Italy
To a first approximation, TrxR inhibition may be attrib-
Supporting information for this article is available on the
WWW under http://dx.doi.org/10.1002/ejic.201300058.
uted to binding of AuI to the C-terminal redox-active –Gly–
Eur. J. Inorg. Chem. 2013, 2718–2727
2718
© 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.eurjic.org
FULL PAPER
Cys–Sec–Gly (GCUG) sequence, as recently suggested by ready at equilibrium once the first spectrum is acquired (i.e.,
MALDI-TOF experiments.[37] However, AuI complexes after few minutes). The equilibrium constant for the reac-
may undergo ligand exchange with several other nucleo- tion with PhSH and PhSeH can be determined by integrat-
philes that are available in living systems, among which ing the acetyl proton signals of auranofin (δ = 1.98, 2.01,
thiol groups are probably the most abundant.[10–12,38–42] Se- 2.05, and 2.08 ppm) and of TATGH (2.01, 2.02, 2.08, and
rum albumin (Alb) is the most abundant plasma protein 2.09 ppm) for different molar ratios of the two reactants
and it contains thiol groups. The interaction of AF with (Figure 2). The equilibrium constant Keq is ca. 1 for PhSH
serum albumin has been widely studied, although a clear and Ͼ 103 for PhSeH. The reaction with PhSeH was investi-
mechanism of interaction is still unknown. In the exchange gated in more detail.
reaction with albumin, the tetraacetylthioglucose portion of
AF is displaced by the Cys-34 thiol group to yield a thiolate
adduct [Alb]S–Au–PEt3, and subsequent release of the
phosphane leads to the formation of Et3PO.[10–12,38–44]
In principle, the formation of such adducts would impede
the delivery of the Au moiety of AF to any other target.
However, selenols and selenides have higher affinities for
AuI than thiols for several factors: (1) the general affinity
between heavy polarizable atoms, (2) the consequent higher
nucleophilic strength of Se than S, and (3) the lower pKa of
selenols (5.2) than thiols (8.0), whereby the –SeH group of
the Sec residue is completely dissociated at physiological
pH, which makes –Se– a better nucleophile than the undis-
sociated thiol.[45–48]
The purpose of this work is to probe the fundamental
features of the thermodynamic stability and relative reactiv-
ity of AuI adducts with S and Se donors (thiols and sele-
nols, respectively) and, thus, provide insights into the
mechanism of action of AuI complexes as TrxR inhibi-
tors.[49] Thus, we wish to understand (1) the reaction path-
ways available to auranofin with thiols, which model its
proposed interaction with albumin, and (2) its relative reac-
tivity with model thiols and selenols, which model the pro-
posed interaction with the enzyme Sec site. However, the
behavior and fate of AuI compounds once they enter the
bloodstream and are delivered to the tissues is beyond the
scope of the present investigation.
To this purpose, we have employed multinuclear NMR
spectroscopy with model thiols and selenols, complemented
by a DFT computational characterization of all species in-
volved. The thiol or selenol functionalities were modeled by
their respective phenyl derivatives, benzenethiol (PhSH) and
benzeneselenol (PhSeH). This choice was motivated by the
1
Figure 2. H NMR (300 MHz, 301 K, CDCl3) spectra of the reac-
tion mixture of auranofin and an equimolar amount of PhSH (top)
and PhSeH (bottom).
The diffusion-ordered (DOSY) spectrum of the 1:1 mix-
need of a selenol less prone to oxidation to diselenide than ture shows that the signals of the Et3P and PhSe moieties
alkyl selenols such as selenocysteine.
have the same diffusion coefficient, D = 1.3ϫ10–9 m2 s–1,
which indicates that they belong to the same molecular en-
tity, whereas D is equal to 1.1ϫ10–9 m2 s–1 for TATGH
(Figure 3).
Results and Discussion
The 31P NMR spectra are hardly informative because the
phosphane resonance at δ = 40.0 ppm is deshielded by less
than 1 ppm upon the addition of up to two equivalents of
Reactions of Auranofin with PhSH and PhSeH in
Chloroform
Upon the addition of one equivalent of PhSH or PhSeH PhSeH.
to a 0.02 m solution of auranofin in CDCl3, displacement of
The 77Se NMR spectrum (at 77 MHz) shows a single res-
the tetraacetylthioglucose moiety by the sulfur or selenium onance at δ = 149.0 ppm (Δν1/2 ≈ 8 Hz), which can be attrib-
nucleophile PhXH (X = S, Se) to form Et3P–Au–XPh takes uted to Et3P–Au–SePh; for comparison, the PhSeH signal
place, as shown by the following observations.
is at δ = 137.6 ppm. After the addition of one further equiv-
1
The H NMR spectra show the rapid appearance of a alent of PhSeH (Figure S1 of the Supporting Information),
doublet at δ = 2.33 ppm (Figure 2) owing to the thiol pro- the sharp resonance (Δν1/2 = 3 Hz) of excess PhSeH appears
ton of tetraacetylthioglucose (TATGH); the reaction is al- at δ = 137.6 ppm, and the other signal is slightly displaced
Eur. J. Inorg. Chem. 2013, 2718–2727
2719
© 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.eurjic.org
FULL PAPER
Figure 3. DOSY spectrum of an equimolar mixture of auranofin and PhSeH in CDCl3.
Scheme 1. Ligand-exchange equilibria of auranofin in chloroform.
to 149.3 ppm and becomes somewhat broader (Δν1/2
18 Hz). As these changes in line width may indicate ex- although it has not been possible to characterize them.
change phenomena, we analyzed the system at lower tem- Therefore, at least in chloroform, which is aprotic and
=
Et3P–Au–SePh are present in solution in rapid equilibrium,
peratures in an attempt to freeze a possible exchange pro- weakly polar, auranofin reacts with the available thiol
cess (Figure S2). The temperature was lowered from 25 to groups and binds reversibly to any thiol donor, but the Au
–60 °C, and the 77Se NMR resonance underwent a linear moiety promptly exchanges with an available selenol func-
upfield shift of ca. 15 ppm. At –60 °C, the 77Se NMR signal tionality. However, such conclusions may not hold in the
was further broadened (Δν1/2 ≈ 70 Hz). Line broadening more polar cell environment or in the aqueous medium em-
caused by relaxation by chemical shift anisotropy was ruled ployed in enzyme inhibition studies. Therefore, we have ex-
out by running the experiment at –60 °C and a lower mag- tended the investigation to a polar and protic solvent such
netic field strength (57 rather than 76 MHz for 77Se), for as methanol, as auranofin is not soluble in water to the
which the line width was still 70 Hz. The dynamic study extent needed for NMR experiments.
could not be brought further owing to solubility problems.
However, the changes in chemical shift and line width indi-
Reactions of Auranofin with PhSH and PhSeH in Methanol
cate that exchange processes operate even at such low tem-
peratures and, therefore, that the exact nature of the adduct
cannot be completely elucidated. Notably, short Au···Au
contacts in the crystal structure of several organophos-
PhSH
1
In Figure 4 we show the H NMR spectra of a 1:1 mix-
phanegold(I) complexes with organic selenolate ligands ture of PhSH and auranofin in CD3OD at different tem-
have been reported.[50,51] Therefore, dimerization and/or peratures.
oligomerization equilibria in solution cannot be excluded.
At low temperature, the resonance of proton H1 of
The overall picture is consistent with the incoming SH TATGH (doublet at δ = 4.76 ppm) is clearly visible. The
or SeH nucleophile rapidly displacing the TATG ligand same signal is at δ = 5.30 ppm in auranofin, and the other
(Scheme 1). Both processes are reversible, but they are char- resonances are almost unaffected. The two resonances co-
acterized by widely different equilibrium constants. How- alesce at ca. 313 K. By applying the Eyring equation with
ever, we cannot exclude that species other than the adduct Δν = 162 Hz, we obtain an exchange frequency kc of 360 Hz
Eur. J. Inorg. Chem. 2013, 2718–2727
2720
© 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.eurjic.org
FULL PAPER
cipitate of gold benzeneselenide develops (see below). Other
decomposition products are also present: only Ph2Se2 was
detected in the 77Se NMR spectrum recorded overnight,
and the 31P NMR spectrum also shows another signal at δ
= 59.2 ppm, which is consistent with Et3PO. Formation of
Et3PO can be understood through further reaction of Et3P–
Au–SePh with PhSeH through Au–P cleavage and subse-
quent formation of [(PhSe)2Au]– and Et3P. [40]
The yellow solid was identified as (C6H5SeAu)n by the
following lines of evidence: (1) Elemental analysis: calcd. C
20.41, H 1.43; found C 20.78, H 1.20. (2) IR spectrum: the
absence of aliphatic C–H stretching below 3000 cm–1 indi-
cates that the triethylphosphane moiety is not present, and
the presence of the phenyl ring is confirmed by arene
stretching bands at 3049, 1471, and 1572 cm–1 and bending
bands at 730 and 695 cm–1. (3) MALDI MS: the spectrum
acquired under negative ionization potential showed three
intense peaks at m/z = 511, 863, and 1217 with an isotopic
pattern consistent with [(PhSeAu)nSePh]– with n = 1, 2, and
3, respectively.
Figure 4. 1H NMR spectra (300 MHz, CD3OD) of a 1:1 mixture
of auranofin and PhSH at (bottom to top) 243, 263, 283, and
313 K. The high-intensity peak that shifts from 5.41 to 4.68 ppm
as the temperature increases is attributed to the hydroxy group.
The reactions between auranofin and PhSeH in CD3OD
are depicted in Scheme 2.
Overall, the reaction of auranofin with a thiol or selenol
can be described as follows. In chloroform, the reaction
leads to clean substitution of the TATG thiolate moiety by
PhX– (X = S, Se) to form a Et3P–Au–XPh adduct. The
reaction is reversible and is characterized by (not surpris-
ingly) a higher affinity of Au for Se rather than S by a factor
of ca. 103 or more. In a more polar solvent such as meth-
anol, the reaction with the model thiol essentially takes the
same course, whereas the reaction with PhSeH is more com-
plex. The initial adduct, in equilibrium with the reactants,
eventually yields Ph2Se2 through the facile oxidation of the
selenol. The adduct then undergoes further reaction with
PhSeH to form [(PhSe)2Au]– and Et3P; the former then po-
lymerizes and the latter is oxidized to Et3PO. Whether these
products are important, or are even formed, in the cell envi-
ronment is an open question.
and an activation free energy of ΔG϶ = 14.7 kcal/mol at
313 K. Therefore, the exchange process is faster in methanol
than in chloroform, in which the signals of both TATGH
and auranofin are well resolved even at room temperature.
In any event, the reaction of auranofin with PhSH in meth-
anol also only amounts to ligand substitution by cleavage
of the Au–S bond.
PhSeH
Immediately after PhSeH (1 equiv.) is mixed with a
0.02 m solution of auranofin in CD3OD, the 1H NMR spec-
trum shows the displacement of TATGH, as in CDCl3 (Fig-
ure S3a), as indicated by the appearance of a doublet at δ
= 4.75 ppm attributed to proton H1 (δ = 5.17 ppm in aur-
anofin). A small shift and broadening of the resonance in
the 31P{1H} NMR spectrum was also observed (less than
1 ppm; Figure S3b), which is consistent with the formation
of Et3P–Au–SePh and again suggests that exchange phe-
Reactions of Et3PAuCl with PhSH and PhSeH in CD3OD
1
nomena occur. No changes were observed in the H NMR
spectra upon lowering of the temperature. However, dis-
To gain more insight into substitution at the gold atom,
placement products are not stable under the reaction condi- we also investigated the reaction of an AuI derivative with
tions, because after 12 h at room temperature a yellow pre- a different leaving group (Et3PAuCl) with the same nucleo-
Scheme 2. Nucleophilic substitution of a sulfur ligand by a selenium ligand in auranofin and further reactions of Et3P–Au–SePh.
Eur. J. Inorg. Chem. 2013, 2718–2727
2721
© 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.eurjic.org
FULL PAPER
philes (PhSH and PhSeH). No reaction occurs in CDCl3. Table 1. Multinuclear NMR spectroscopic data (ppm).
Therefore, the process was investigated in CD3OD by add-
ing one equivalent of PhXH to a 0.02 m solution of Et3-
PAuCl.
Species
1H
31P 77Se
35Cl
AF
(CDCl3)
AF
PhSeH
TATGH
ca. 5.14 (H1) 40.0
–
2.31 (SH),
4.54 (H1)
–
–
–
–
–
–
–
137.6
–
PhSH
Immediately after the addition of PhSH, the 31P NMR
resonance shifts from δ = 33.6 to 35.5 ppm. After one day,
we only observed the formation of [(Et3P)2Au]+, as indi-
cated by the 31P resonance at δ = 47.4 ppm (lit.: 45.7[40,52]
and 47.25 ppm[53] depending on the counterion) along with
a white precipitate.
PhSeAu-
PEt3
PhSeSePh
40.8 149.0[a]
–
–
–
–
457.3
AF
(MeOD)
AF
5.17 (H1)
–
4.75 (H1)
–
38.85 –
–
–
–
–
PhSeH
TATGH
PhSeAu-
PEt3
–
–
106.4
–
39.85 123.4[a]
PhSeH
PhSeSePh
Et3P=O
–
–
–
59.2
452.3
–
–
–
Immediately after the addition of PhSeH, the 31P{1H}
NMR spectrum (Figure 5) shows a new intense peak at δ =
37.2 ppm, which is 3.6 ppm more deshielded than the 31P
NMR resonance of the reactant. This signal is attributed to
Et3P–Au–SePh; however, this value does not match the one
for the same putative complex obtained from auranofin (ca.
40 ppm; Table 1), which probably indicates that some ex-
change process between Et3P–Au–SePh and the reactant
occurs. However, after three days several new 31P signals
Et3PAuCl
(MeOD)
Et3PAuCl
PhSeAu-
PEt3
PhSAuPEt3
(Et3P)2AuCl
Et3PD+
Cl–
–
–
33.6
37.2 123.4,
–
–
–
185.4[a]
–
–
–
–
35.5
47.1
22.0
–
–
–
–
–
–
–
–
–32.7
[a] As discussed in the text, more than one species in rapid ex-
change may contribute to this signal.
can be observed. (1) A weak 1:1:1 triplet at δ = 22.2 ppm,
1
1
assigned to Et3PD+ with J
= 73 Hz. (2) A resonance
(δ = 106.4 ppm in MeOD). The H NMR spectra, except
31P,2H
at δ = 47.1 ppm owing to (Et3P)2AuCl.[53] (3) A resonance
at δ = 59.7 ppm, which approximately matches that of
Et3PO,[40] except that it also features two satellite peaks (ca.
3.5% each) separated by 424 Hz, which suggests the forma-
tion of an unidentified compound with a P–Se bond.[54]
for the evidence of the formation of Ph2Se2, were not in-
formative as the signals of all Et3P groups largely over-
lapped. It is apparent that oxygen plays a key role in the
promotion of the oxidation of Et3PD+ to Et3PO.
Figure 6. 77Se{1H} spectrum (57 MHz, CD3OD, 301 K) of the mix-
ture Et3PAuCl/PhSeH under N2 after three days.
The displacement of chloride is borne out by 35Cl NMR
spectroscopy. The spectral lines of covalent molecules with
monovalent Cl have line widths in the order of 10 kHz and
are undetectable for all practical purposes, whereas the 35Cl
nucleus in the Cl– ion experiences a vanishing electric field
gradient and gives rise to correspondingly narrow lines.[55]
Indeed, the displacement of Cl– is clearly evident in the 35Cl
NMR spectrum of the reaction mixture, in which a reso-
nance at δ = –32.7 ppm is observed (Figure 7).
Figure 5. 31P{1H} NMR spectra (122 MHz, CD3OD, 301 K) of
(from bottom to top): Et3PAuCl, Et3PAuCl + PhSeH (1:1) just after
mixing, Et3PAuCl + PhSeH (1:1) after three days, and Et3PAuCl +
PhSeH (1:1) after three days under a N2 atmosphere.
In addition to the reactions mentioned above, the pre-
cipitation of a yellow solid also occurs, which was identified
If the reaction is run under a N2 atmosphere (Figure 5, as (PhSeAu)n as before (vide supra).
top trace), most of these additional products do not form
Therefore, it seems that the reaction of Et3PAuCl with
and the signal from Et3PD+ is stronger. Moreover, under an both PhSH and PhSeH features initial coordination of a
inert atmosphere, the 77Se{1H} NMR spectrum (Figure 6) S or Se atom to the Au center, followed by the release of
shows three signals: Ph2Se2 (δ = 452.3 ppm), Et3P–Au–SePh triethylphosphonium ion (Et3PH+), that is, cleavage of the
(δ = 123.4 ppm, broad), and an unassigned resonance at δ Au–P bond. The remaining gold(I) benzeneselenide (or
= 185.3 ppm. None of these signals are attributed to PhSeH -sulfide) precipitates. Excess triethylphosphane partly coor-
Eur. J. Inorg. Chem. 2013, 2718–2727
2722
© 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.eurjic.org
FULL PAPER
Figure 7. 35Cl NMR spectrum (29 MHz, CD3OD, 28 °C) of Cl– displaced by the reaction between Et3PAuCl and PhSeH.
dinates to Et3PAuCl, which leads to the formation of which we have replaced the TATG, phenyl, and ethyl moie-
(Et3P)2AuCl. Other reactions occur with PhSeH; however, ties with smaller methyl groups and adopted CH3SH or
essentially all selenium species eventually undergoes oxi- CH3SeH as model thiol and selenol for simplicity.
dation to Ph2Se2 (which is the only species observed in the
77Se spectrum).
As with auranofin, under both air and N2, the 31P NMR
resonance of the reactant is deshielded in the initial stages
of the reaction and then returns to the initial value. This
suggests the overall reaction scheme shown in Scheme 3.
The reaction with CH3SH [Reaction (1)] is of course an
identity reaction with ΔE and ΔG = 0. Reactions (2a) and
(2b) model the two conceivable reactions of auranofin with
PhSeH through Au–S and Au–P cleavage, respectively, and
Reactions (3a) and (3b) are the analogous reactions of
(CH3)3P–Au–Cl. Reactions (2a) and (3a) also concern dis-
placement of the ligand trans to the phosphane, whereas
Scheme 3. Reaction of Et3PAuCl with a) PhSH and b) PhSeH in
Reactions (2b) and (3b) concern the displacement of tri-
polar solvents. After one day a precipitate appears (white or yellow
methylphosphane, which leads to an ion-pair product.
for PhSH and PhSeH, respectively). In panel b) we have highlighted
First, we have characterized the thermodynamics of all
processes through the electronic energy (ΔE) and the free
energy (ΔG) changes. The results are reported in Table 2 for
the reactions in gas phase and in aqueous solution.
the deuterium exchange with the solvent as we detected the phos-
phonium ion Et3PD+.
The experimental results presented above indicate that,
in general, ligand substitution at the gold center firstly takes
place at the ligand trans to the phosphane (S or Se), as
generally accepted. However, the formation of triethylphos-
phane and derivatives indicates that the Au–P bond may
also be cleaved under certain conditions. In all cases, the
mechanism for substitution, by either displacement or ad-
dition/elimination, is also unknown. These issues have been
clarified with recourse to the calculation of the potential
energy surface of model systems, including the characteriza-
tion of the relevant transition states.
Table 2. Thermodynamics [kcal/mol] in the gas and aqueous phases
of ligand substitution at gold in Me3PAuX.
Reaction
Gas phase
ΔEg
Aqueous solution
ΔGg
ΔEs
ΔGs
ΔGne
1
0.0
0.0
0.0
–10.1
4.1
13.3
6.1
0.0
0.0
2a
2b
3a
3b
–10.2
92.1
–1.0
90.4
–10.2
90.3
1.7
–10.2
3.8
–0.7
–0.5
–0.5
–0.1
–2.1[a]
5.0
89.7
DFT Calculations
[a] The energy was obtained by considering the dissociated species
H3O+ and Cl–. The solvation free energy of H3O+ was taken from
ref.[56] See also ref.[57] for the calculation of the solvation free energy
The reactions investigated feature a simplified model
gold complex, namely, (CH3)3P–Au–X (X = SCH3, Cl), in of H3O+.
Eur. J. Inorg. Chem. 2013, 2718–2727
2723
© 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.eurjic.org
FULL PAPER
Reactions that involve only neutral reactants and prod- directly transferred from the Se atom to the S atom through
ucts show very small solvent effects, as expected; conversely, the transition state TS1. The reaction coordinate indicated
Reactions (2b) and (3b), in which ion pairs are formed, do by the normal mode with an imaginary frequency is a mix-
show large effects and the gas-phase results can be essen- ture of Se–H stretching and Se–H–S bending, and the struc-
tially ignored. The large negative free energy of Reac- ture is closer to the reactant than to the products, in agree-
tion (2a), the displacement of a S ligand by a Se ligand, ment with the Hammond postulate for an exothermic reac-
both in the gas phase and in solution (ΔGg = ΔGs
=
tion. For comparison we have calculated the analogous
–10.2 kcal/mol) indicates a marked preference of the gold transition state for the symmetric Reaction (1) and for the
centre to coordinate with a selenium ligand rather than a endoergonic Reaction (3a), see Figure S4. In the second
sulfur ligand, which is consistent with previous calcula- path, the proton transfer is mediated by the gold centre,
tions[58] and the experimental findings described above. The through the formation of an intermediate I with a square
displacement of a phosphane ligand by a Se ligand [Reac- tetracoordinate hydride gold complex (Figure S5). Because
tions (2b) and (3b)] is thermodynamically much less favor- of the high energy and unusual structure of the intermedi-
able for both (CH3)3P–Au–SCH3 and (CH3)3P–Au–Cl. We ate, this path can be safely discarded.
also note that nonelectrostatic contributions to the sol-
All the structures involved in Reaction (2a) have been re-
vation free energy (last column in Table 2) are rather small. optimized by using the polarizable continuum model
Hereafter, we will present the computational characteri- (PCM) method to investigate the reaction mechanism in
zation of the reaction pathways available for the reactions water. The structures undergo minor changes; the energy
of Table 2; however, as some general mechanisms are very profile also changes slightly, but the general features of both
high in energy, the description of the involved molecular reaction paths found in the gas phase are unaltered. The
structures is deferred to the Supporting Information (Fig- relative free energy of all characterized structures in the
ures S4–S7).
presence of the solvent are reported in Table S1 together
with the corresponding values in the gas phase for compari-
son.
Mechanism of Reactions Involving Au–S Cleavage
For Reaction (2a), we have completely characterized two
alternative paths (Figure 8), in which the Gibbs energy of
the transition states and intermediates are also indicated
(the free energy of the van der Waals complex of the reac-
tants has been chosen as reference). We have indicated the
van der Waals complexes of the reactants and products as
R and P, respectively.
Mechanism of Reaction Involving Au–P Cleavage
Reaction (2b) involves the formation of two charged
products from two neutral species and, as expected, it is
highly endoergonic in the gas phase (ΔGg = 90.3 kcal/mol).
Two reaction paths have been characterized (Figure 9).
Figure 9. Top: schematic representation of the free energy reaction
paths for Reaction (2b) in the gas phase. Bottom: 3D structures of
reactants, TS, and products of the low-energy path.
Figure 8. Top: schematic representation of the two free energy reac-
tion paths for Reaction (2a) in the gas phase. Bottom: 3D struc-
tures of reactants, TS, and products of the low-energy path.
Two almost isoenergetic (ΔG Ͻ 2 kcal/mol) structures of
reactants and products have been obtained: for the van der
The two paths differ in the mechanism of hydrogen trans- Waals complex of the reactants the Se–H bond can be di-
fer: in the lowest energy path, the selenol hydrogen atom is rected toward the sulfur atom or the gold atom. Also for
Eur. J. Inorg. Chem. 2013, 2718–2727
2724
© 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.eurjic.org
FULL PAPER
the products, which for Reaction (2b) are in the form of an reactions are triggered by oxidation of PhSH and PhSeH
ion pair between phosphonium and the gold complex, the and displacement of triethylphosphane. As a result, a com-
acidic hydrogen atom of the phosphonium can be directed plex mixture of equilibrating products is observed, which
toward the gold center or toward the Se atom. In the reac- could not be fully characterized as exchanges among them
tion paths of Figure 9, we have considered the most stable are fast on the NMR timescale. An important result con-
arrangements, that is, the one with the hydrogen atom cerns the reversibility of the reaction with thiols: the bio-
pointing toward the sulfur atom for the reactants and the chemical implication is that even after binding to albumin
one with the hydrogen atom pointing toward the selenium in the blood, the Et3PAu moiety of auranofin can be easily
atom in the products.
exchanged with sites containing selenium atoms such as the
In the lower energy path, the transition state (TS1) corre- selenocysteine residue in TrxR, which ensures its efficient
sponds to the almost complete formation of the Se–Au delivery. The reaction paths of the two reactions at the gold
bond (the bond length in the TS1 is only 0.1 Å longer than center have been investigated computationally; a concerted
in the products) and the concurrent displacement of the mechanism, in which the proton of the incoming nucleo-
phosphane moiety. The imaginary normal mode involves phile (PhSeH or PhSH) is attached to the leaving group and
P–Au stretching together with bending of the Se–Au–S the new chalcogenide–Au bond is formed, has been found
bond. The relatively high activation energy of such a path to be the favored mechanism for the ligand exchange pro-
qualitatively agrees with the experimental evidence of a fac- cess. In agreement with the experimental observations, we
ile displacement of the thiol ligand rather than of the phos- have found that the substitution of –SCH3 by –SeCH3 is
phane from the gold centre.
favored. However, calculations including the solvent reac-
In TS1, the phosphane leaves the complex in its neutral tion field indicate that displacement of the phosphane may
form P(CH3)3, and the hydrogen atom remains bound to also occur.
the Se atom. For the reaction in the gas phase, we also find
the transition state (TSH) which corresponds to the direct
transfer of the proton from the Se atom to the phosphane
Experimental Section
to form the phosphonium ion (Figure S7).
Sample Preparation and Characterization: Auranofin (Vinci-Bio-
chem), triethylphosphane gold(I) chloride (Et3PAuCl), benzene-
thiol (PhSH), and benzeneselenol (PhSeH, Aldrich) were used
without further purification. Solutions (ca. 0.02 m) were prepared
directly in NMR tubes with 0.6 mL of deuterated solvent.
The structures involved in this mechanism have also been
optimized in water. As before, the geometries undergo
minor changes. Moreover, for the reaction in aqueous solu-
tion the hydrogen exchange, which in the gas phase is de-
scribed by TSH, most probably occurs with the solvent and,
thus, no direct proton exchange between the Se atom and
the phosphane is expected. This is confirmed by the
31P{1H} NMR spectrum of the reaction in deuterated
methanol, which shows a triplet consistent with the ex-
FTIR Spectroscopy: IR Spectra were obtained with a Nicolet 5700
instrument equipped with a Germanium ATR Thermo Smart Per-
former accessory.
MS Spectrometry: LDI/MS spectra were obtained with an Ul-
traflex II MALDI-TOF instrument (Bruker Daltonics) operating
in reflection negative ion mode. Ions were formed by a pulsed UV
laser (λ = 337 nm) beam. The instrumental conditions were: IS1 =
20 kV, IS2 = 17.3 kV, reflection potential = 21 kV, delay time 70 ns.
1 μL of sample solution was deposited on the stainless steel sample
holder and allowed to dry before introduction into the mass spec-
trometer. External mass calibration (peptide calibration standard)
was based on monoisotopic values of [M – H]– of 2,5-dhydroxyben-
zoic acid, sinapinic acid, angiotensin II, angiotensin I, substance P,
bombesin, ACTH clip (1–17), ACTH clip (18–39), somatostatin 28
at m/z = 153.020, 223.062, 1044.527, 1294.670, 1345.720, 1617.808,
2091.072, 2463.184, and 3145.456, respectively.
+
change of –PD3 with the deuterated solvent. In solution,
the final ion pair is stabilized by about 21.5 kcal/mol with
respect to the gas phase, and the gold complex attains an
almost perfect linear geometry (S–Au–Se 179.4°). The rela-
tive free energy of the optimized structures in the presence
of the solvent are reported in Table S2 together with the
corresponding values of the reaction in the gas phase for
comparison.
The second, higher-energy path is similar to that found
for Reaction (2a), in which the hydrogen transfer is medi-
ated by the gold centre. The optimized structures of the two
transition states and the reaction intermediate are shown in
Figure S7.
NMR Spectroscopy: NMR spectra, including DOSY, were acquired
with a Bruker AVANCE 300 spectrometer equipped with a 5-mm
BBO z-grad probe with a maximum gradient strength of 53 G/cm.
Operating frequencies are 300 (1H), 75 (13C), 122 (31P), 57 (77Se),
and 29 MHz (35Cl). Chemical shifts are reported with respect to
1
the following reference standards: tetramethylsilane (TMS; H and
13C), 85% aqueous H3PO4 (31P), Me2Se (indirectly calibrated by
setting the signal of a 4.4 m solution of selenophene in CDCl3 to
605 ppm; 77Se),[55] and 0.1 m NaCl in D2O (35Cl). Low temperature
77Se NMR spectra (76 MHz) shown in the Supporting Information
were acquired with an AVANCE DRX 400 spectrometer equipped
with a 5-mm BBI z-grad probe. 13C, 31P, and 77Se spectra were
proton-decoupled.
Conclusions
We have investigated the reaction of model sulfur and
selenium nucleophiles (PhSH and PhSeH) with auranofin
and Et3PAuCl. We have highlighted a clear difference in the
stability of the adducts formed by S and Se derivatives, and
the latter are more stable. Nevertheless, in addition to the
adduct Et3PAu–X–Ph, other species are also formed; this is
DFT Calculations: Calculations were run with the Gaussian 03 pro-
particularly prominent in polar solvents, in which further gram package.[59] The B3PW91 hybrid functional[60,61] was em-
Eur. J. Inorg. Chem. 2013, 2718–2727
2725
© 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.eurjic.org
FULL PAPER
ployed in combination with the LANL2DZ basis set[62–64] for Au
and Se and the standard 6-31G** basis set[65] for the other atoms.
This combination has been shown to provide satisfactory results
for thiolate and phosphane complexes of Group 11 metals.[58,66–68]
Normal coordinate analysis was used to determine the imaginary
frequencies of transition states structures, and no imaginary fre-
quencies were found in all other cases. For each transition structure
we calculated the intrinsic reaction coordinate (IRC) for six points
towards the corresponding minima in both directions (reverse and
forward) and then we performed a geometry optimization from the
final point of the IRC path towards the minima.
[12]
A. P. Intoccia, T. L. Flanagan, D. T. Walz, l. Gutzait, J. E.
Swagzdis, K. J. Flagiello, B. Y. Hwang, R. H. Dewey, H. Nog-
ouchi, in: Bioinorganic Chemistry of Gold Coordination Com-
pounds, Smith Kline and French Laboratories, Philadelphia,
1983, pp. 21–33.
L. R. Kelland, Crit. Rev. Oncol. Hematol. 1993, 15, 191–219.
S. Gromer, S. Urig, K. Becker, Med. Res. Rev. 2004, 24, 40–89.
E. S. J. Arner, Biochim. Biophys. Acta Gen. Subj. 2009, 1790,
495–526.
C. Marzano, V. Gandin, M. Pellei, D. Colavito, G. Papini,
G. G. Lobbia, E. Del Giudice, M. Porchia, F. Tisato, C. San-
tini, J. Med. Chem. 2008, 51, 798–808.
V. Gandin, A. P. Fernandes, M. P. Rigobello, B. Dani, F. Sor-
rentino, F. Tisato, M. Bjornstedt, A. Bindoli, A. Sturaro, R.
Rella, C. Marzano, Biochem. Pharmacol. 2010, 79, 90–101.
C. Marzano, M. Pellei, D. Colavito, S. Alidori, G. G. Lobbia,
V. Gandin, F. Tisato, C. Santini, J. Med. Chem. 2006, 49, 7317–
7324.
M. Rigobello, A. Folda, G. Scutari, A. Bindoli, Arch. Biochem.
Biophys. 2005, 441, 112–122.
M. Gimeno, A. Laguna, Chem. Rev. 1997, 97, 511–522.
B. M. Sutton, E. McGusty, D. T. Walz, M. J. DiMartino, J.
Med. Chem. 1972, 15, 1095–1098.
A. J. Lewis, D. T. Walz, Immunopharmacology of Gold, in: Pro-
gress in Medicinal Chemistry, vol. 19 (Eds.: G. P. Ellis, G. B.
West), Elsevier, 1982, pp. 1–58.
A. E. Finkelstein, D. T. Walz, V. Batista, M. Mizraji, F. Rois-
man, A. Misher, Ann. Rheum. Dis. 1976, 35, 251–257.
A. Glennas, T. Kvien, O. Andrup, O. ClarkeJenssen, B. Kar-
stensen, U. Brodin, Br. J. Rheumatol. 1997, 36, 870–877.
E. J. Brewer Jr., E. H. Giannini, D. A. Person, Am. J. Med.
1983, 75, 152–156.
[13]
[14]
[15]
[16]
[17]
[18]
[19]
All structures were firstly optimized in the gas phase. The effect of
a polar solvent on the structures was then modeled by reoptimizing
the structures by means of the PCM[69–73] as implemented in
Gaussian 03.[59] Water was chosen as the solvent chosen for PCM
calculations (ε = 78.39).
The standard free energies of reactions in the gas phase and in
solution are denoted as ΔG0 and ΔG0 , respectively. From Hess’s
g
s
law, ΔG0 = ΔG0 + ΔG0solv, in which ΔG0
is the difference be-
[20]
[21]
s
g
solv
tween the free energy of solvation of the products and the reactants.
Within the PCM, the free energy of solvation is composed of the
electrostatic (ΔGel) and nonelectrostatic (ΔGne) contributions
(ΔGsolv = ΔGel + ΔGne). In turn, ΔGne = ΔGdis + ΔGrep + ΔGcav, in
which ΔGdis and ΔGrep are the solvent–solute dispersion and repul-
sion interactions, respectively, and ΔGcav is the cavitation energy.
We report ΔG0s as obtained from the results of the thermochemical
analysis of the optimized solvated structures. Such values do not
include the nonelectrostatic part of ΔGsolv, which are reported sepa-
rately and are negligible for the considered reactions in water
(ΔGne Ͻ 1 kcal/mol). Images of reaction intermediates and transi-
tion states were produced with Jmol.[74]
[22]
[23]
[24]
[25]
[26]
[27]
C. F. Shaw, A. A. Isab, J. D. Hoeschele, M. Starich, J. Locke,
P. Schulteis, J. Xiao, J. Am. Chem. Soc. 1994, 116, 2254–2260.
J. L. Hickey, R. A. Ruhayel, P. J. Barnard, M. V. Baker, S.
Berners-Price, A. Filipovska, J. Am. Chem. Soc. 2008, 130,
12570–12571.
Supporting Information (see footnote on the first page of this arti-
cle): Additional NMR spectra and computational data.
[28]
[29]
[30]
[31]
S. J. Berners-Price, M. J. Dimartino, D. T. Hill, R. Kuroda,
M. A. Maziz, P. J. Sadler, Inorg. Chem. 1985, 24, 3425–3434.
D. T. Walz, M. J. Di Martino, C. L. Griffin, A. Misher, Arch.
Int. Pharmacodyn. Ther. 1970, 85, 337–343.
M. Rigobello, G. Scutari, R. Boscolo, A. Bindoli, Br. J. Pharm-
acol. 2002, 136, 1162–1168.
S. Gromer, L. Arscott, C. Williams, R. Schirmer, K. Becker, J.
Biol. Chem. 1998, 273, 20096–20101.
M. McKeage, Br. J. Pharmacol. 2002, 136, 1081–1082.
M. Rigobello, G. Scutari, A. Folda, A. Bindoli, Biochem. Phar-
macol. 2004, 67, 689–696.
Acknowledgments
We thank Fondazione CARIPARO for a Ph. D. fellowship to FDS,
F. Tisato (CNR, Padova) for MALDI MS spectra, C. Marzano
and V. Gandin (University of Padova) for useful discussions and
F. Rastrelli and I. Menegazzo (University of Padova) for help with
NMR experiments.
[32]
[33]
[1] D. T. Walz, D. E. Griswold, Inflammation 1978, 3, 117–128.
[2] Y. Hamada, F. Shinomiya, M. Okada, T. Fujimura, Mod.
Rheumatol. 2003, 13, 0027–0034.
[34]
[35]
D. Green, Cell 1998, 94, 695–698.
I. Kim, J. Jin, I. Lee, S. Park, Br. J. Pharmacol. 2004, 142, 749–
755.
[3] S. J. Berners-Price, L. Ronconi, P. J. Sadler, Prog. Nucl. Magn.
Reson. Spectrosc. 2006, 49, 65–98.
[36]
[37]
S. Park, I. Kim, Br. J. Pharmacol. 2005, 146, 506–513.
C. Gabbiani, G. Mastrobuoni, F. Sorrentino, B. Dani, M. P.
Rigobello, A. Bindoli, M. A. Cinellu, G. Pieraccini, L. Messori,
A. Casini, Med. Chem. Commun. 2011, 2, 50–54.
C. F. Shaw, N. A. Schaeffer, R. C. Elder, M. K. Eidsness, J. M.
Trooster, G. H. M. Calis, J. Am. Chem. Soc. 1984, 106, 3511–
3521.
A. A. Isab, A. L. Hormann, D. T. Hill, D. E. Griswold, M. J.
Dimartino, C. F. Shaw, Inorg. Chem. 1989, 28, 1321–1326.
A. A. Isab, C. F. Shaw, J. D. Hoeschele, J. Locke, Inorg. Chem.
1988, 27, 3588–3592.
N. A. Malik, G. Otiko, M. T. Razi, P. J. Sadler, B. M. Sutton,
R. G. Franz in Bioinorganic Chemistry of Gold Coordination
Compounds, Smith Kline and French Laboratories, Philadel-
phia, 1983, pp. 82–97.
M. T. Coffer, C. F. Shaw, M. K. Eidsness, J. W. Watkins, R. C.
Elder, Inorg. Chem. 1986, 25, 333–339.
[4] G. Natile, L. G. Marzilli, Coord. Chem. Rev. 2006, 250, 1315–
1331.
[5] C. H. Williams, Eur. J. Biochem. 2000, 267, 6101–6101.
[6] T. Sandalova, L. W. Zhong, Y. Lindqvist, A. Holmgren, G.
Schneider, Proc. Natl. Acad. Sci. USA 2001, 98, 9533–9538.
[7] E. I. Biterova, A. A. Turanov, V. N. Gladyshev, J. J. Barycki,
Proc. Natl. Acad. Sci. USA 2005, 102, 15018–15023.
[8] K. Becker, S. Gromer, R. H. Schirmer, S. Muller, Eur. J. Bio-
chem. 2000, 267, 6118–6125.
[9] E. H. Giannini, K. S. Barron, C. H. Spencer, D. A. Person, J.
Baum, B. H. Bernstein, D. W. Kredich, J. C. Jacobs, L. S.
Zemel, D. Gibbas, M. H. Passo, E. J. Brewer, J. Rheumatol.
1991, 18, 1240–1242.
[38]
[39]
[40]
[41]
[10] M. T. Coffer, C. F. Shaw, A. L. Hormann, C. K. Mirabelli, S. T.
Crooke, J. Inorg. Biochem. 1987, 30, 177–187.
[11] N. A. Malik, G. Otiko, P. J. Sadler, J. Inorg. Biochem. 1980, 12,
317–322.
[42]
Eur. J. Inorg. Chem. 2013, 2718–2727
2726
© 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.eurjic.org
FULL PAPER
[43] L. Messori, A. Balerna, I. Ascone, C. Castellano, C. Gabbiani,
A. Casini, C. Marchioni, G. Jaouen, A. C. Castellano, J. Biol.
Inorg. Chem. 2011, 16, 491–499.
[44] J. Roberts, J. Xiao, B. Schleisman, D. Parsons, C. Shaw, Inorg.
Chem. 1996, 35, 424–433.
[45] A. P. Arnold, K. S. Tan, D. L. Rabenstein, Inorg. Chem. 1986,
25, 2433–2437.
[46] A. A. Isab, Transit. Met. Chem. 1994, 19, 595–598.
[47] S. D. Lewis, D. C. Misra, J. A. Shafer, Biochemistry 1980, 19,
6129–6137.
Voth, P. Salvador, J. J. Dannenberg, V. G. Zakrzewski, S. Dap-
prich, A. D. Daniels, M. C. Strain, O. Farkas, D. K. Malick,
A. D. Rabuck, K. Raghavachari, J. B. Foresman, J. V. Ortiz, Q.
Cui, A. G. Baboul, S. Clifford, J. Cioslowski, B. B. Stefanov,
G. Liu, A. Liashenko, P. Piskorz, I. Komaromi, R. L. Martin,
D. J. Fox, T. Keith, M. A. Al-Laham, C. Y. Peng, A. Nanayak-
kara, M. Challacombe, P. M. W. Gill, B. Johnson, W. Chen,
M. W. Wong, C. Gonzalez, J. A. Pople, Gaussian 03, rev. D.02,
Gaussian, Inc., Wallingford CT, 2003.
[60] A. D. Becke, J. Chem. Phys. 1993, 98, 5648–5652.
[61] J. P. Perdew, Electronic Structure of Solids ’91, Akademie Ver-
lag, Berlin.
[48] J. C. Pleasants, W. Guo, D. L. Rabenstein, J. Am. Chem. Soc.
1989, 111, 6553–6558.
[49] S. Urig, K. Becker, Semin. Cancer Biol. 2006, 16, 452–465.
[50] P. G. Jones, C. Thöne, Chem. Ber. 1990, 123, 1975–1978.
[51] W. Eikens, C. Kienitz, P. G. Jones, C. Thone, J. Chem. Soc.,
Dalton Trans. 1994, 83–90.
[52] A. A. Isab, A. L. Hormann, M. T. Coffer, C. F. Shaw, J. Am.
Chem. Soc. 1988, 110, 3278–3284.
[62] P. J. Hay, W. R. Wadt, J. Chem. Phys. 1985, 82, 270–283.
[63] P. J. Hay, W. R. Wadt, J. Chem. Phys. 1985, 82, 299–310.
[64] W. R. Wadt, P. J. Hay, J. Chem. Phys. 1985, 82, 284–298.
[65] W. J. Hehre, R. Ditchfield, J. A. Pople, J. Chem. Phys. 1972, 56,
2257–2261.
[66] J. A. S. Howell, Polyhedron 2006, 25, 2993–3005.
[67] P. Pyykko, Angew. Chem. 2004, 116, 4512; Angew. Chem. Int.
Ed. 2004, 43, 4412–4456.
[53] A. Al Arfaj, A. Saeed, M. Akhtar, A. Isab, J. Coord. Chem.
1998, 43, 257–272.
[54] K. P. Bhabak, G. Mugesh, Inorg. Chem. 2009, 48, 2449–2455.
[55] J. Mason, Multinuclear NMR, Plenum Press, New York, 1987.
[56] R. M. Noyes, J. Am. Chem. Soc. 1962, 84, 513–522.
[57] G. Saielli, J. Phys. Chem. A 2010, 114, 7261–7265.
[58] J. A. S. Howell, J. Organomet. Chem. 2009, 694, 868–873.
[59] M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria,
M. A. Robb, J. R. Cheeseman, J. J. A. Montgomery, T. Vreven,
K. N. Kudin, J. C. Burant, J. M. Millam, S. S. Iyengar, J. Tom-
asi, V. Barone, B. Mennucci, M. Cossi, G. Scalmani, N. Rega,
G. A. Petersson, H. Nakatsuji, M. Hada, M. Ehara, K. Toyota,
R. Fukuda, J. Hasegawa, M. Ishida, T. Nakajima, Y. Honda,
O. Kitao, H. Nakai, M. Klene, X. Li, J. E. Knox, H. P. Hratch-
ian, J. B. Cross, V. Bakken, C. Adamo, J. Jaramillo, R. Gom-
perts, R. E. Stratmann, O. Yazyev, A. J. Austin, R. Cammi, C.
Pomelli, J. W. Ochterski, P. Y. Ayala, K. Morokuma, G. A.
[68] G. S. M. Tong, A. S. C. Cheung, J. Phys. Chem. A 2002, 106,
11637–11643.
[69] M. Cossi, V. Barone, R. Cammi, J. Tomasi, Chem. Phys. Lett.
1996, 255, 327–335.
[70] M. Cossi, N. Rega, G. Scalmani, V. Barone, J. Chem. Phys.
2001, 114, 5691–5701.
[71] M. Cossi, G. Scalmani, N. Rega, V. Barone, J. Chem. Phys.
2002, 117, 43–54.
[72] E. Cances, B. Mennucci, J. Tomasi, J. Chem. Phys. 1997, 107,
3032–3041.
[73] J. Tomasi, B. Mennucci, E. Cances, THEOCHEM 1999, 464,
211–226.
[74] R. M. Hanson, J. Appl. Crystallogr. 2010, 43, 1250–1260.
Received: January 16, 2013
Published Online: April 2, 2013
Eur. J. Inorg. Chem. 2013, 2718–2727
2727
© 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim