See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/232969231

Protein metalation by metal-based drugs: Reactions of cytotoxic gold

compounds with cytochrome c and lysozyme

Article in European Journal of Biochemistry · November 2012

DOI: 10.1007/s00775-012-0952-6 · Source: PubMed

CITATIONS READS

39 262

7 authors, including:

Chiara Gabbiani Lara Massai

Università di Pisa University of Florence
138 PUBLICATIONS 5,617 CITATIONS 59 PUBLICATIONS 946 CITATIONS

SEE PROFILE SEE PROFILE

Elena Michelucci Laura Maiore

Italian National Research Council Università degli Studi di Sassari
48 PUBLICATIONS 1,020 CITATIONS 27 PUBLICATIONS 555 CITATIONS

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Electrochemical and spectroelectochemical behavior of organometallic compounds in configuration d8 View project

Arsenoplatin-1 is a Dual Pharmacophore Anti-Cancer Agent View project

All content following this page was uploaded by Laura Maiore on 19 May 2014.

The user has requested enhancement of the downloaded file.

J Biol Inorg Chem (2012) 17:1293–1302
DOI 10.1007/s00775-012-0952-6
ORIGINAL PAPER
Protein metalation by metal-based drugs: reactions of cytotoxic
gold compounds with cytochrome c and lysozyme
Chiara Gabbiani • Lara Massai • Federica Scaletti •
Elena Michelucci • Laura Maiore • Maria Agostina Cinellu
Luigi Messori
Received: 6 August 2012 / Accepted: 7 October 2012 / Published online: 7 November 2012
Ó SBIC 2012
Abstract Protein metalation processes are crucial for the
mechanism of action of several anticancer metallodrugs
and warrant deeper characterisation. We have explored the
reactions of three cytotoxic gold(III) compounds—namely
[(bipy2Me)2Au2(l-O)2][PF6]2 (where bipy2Me is 6,60 -dime-
thyl-2,20 -bipyridine) (Auoxo6), [(phen2Me)2Au2(l-O)2]
[PF6]2 (where phen2Me is 2,9-dimethyl-1,10-phenanthro-
line) (Au2phen) and [(bipydmb-H)Au(OH)][PF6] [where
bipydmb-H is deprotonated 6-(1,1-dimethylbenzyl)-2,20 -
bipyridine] (Aubipyc)—with two representative model
proteins, i.e. horse heart cytochrome c and hen egg white
Electronic supplementary material The online version of this
article (doi:10.1007/s00775-012-0952-6) contains supplementary
material, which is available to authorized users.
C. Gabbiani
Department of Chemistry and Industrial Chemistry,
University of Pisa,
Via Risorgimento 35,
56126 Pisa, Italy
L. Massai  F. Scaletti  L. Messori (&)
Department of Chemistry,
University of Florence,
Via della Lastruccia 3,
50019 Sesto Fiorentino, Italy
e-mail: luigi.messori@unifi.it
E. Michelucci
Mass Spectrometry Centre (CISM),
University of Florence,
Via U. Schiff 6,
50019 Sesto Fiorentino, Italy
L. Maiore  M. A. Cinellu
Department of Chemistry and Pharmacy,
University of Sassari,
Via Vienna 2,
07100 Sassari, Italy
•
lysozyme, through UV–visible absorption spectroscopy
and electrospray ionisation mass spectrometry (ESI MS) to
characterise the inherent protein metalation processes.
Notably, Auoxo6 and Au2phen produced stable protein
adducts where one or more ‘‘naked’’ gold(I) ions are pro-
tein-coordinated; very characteristic is the case of cyto-
chrome c, which upon reaction with Auoxo6 or Au2phen
preferentially forms ‘‘tetragold’’ adducts with four protein-
bound gold(I) ions. In turn, Aubipyc afforded monometa-
lated protein adducts where the structural core of the
gold(III) centre and its ?3 oxidation state are conserved.
Auranofin yielded protein derivatives containing the intact
auranofin molecule. Additional studies were performed to
assess the role played by a reducing environment in protein
metalation. Overall, the approach adopted provides
detailed insight into the formation of metallodrug–protein
derivatives and permits trends, peculiarities and mecha-
nistic details of the underlying processes to be highlighted.
In this respect, electrospray ionisation mass spectrometry is
a very straightforward and informative research tool. The
protein metalation processes investigated critically depend
on the nature of both the metal compound and the inter-
acting protein and also on the solution conditions used;
thus, predicting with accuracy the nature and the amounts
of the adducts formed for a given metallodrug–protein pair
is currently extremely difficult.
Keywords Anticancer drugs  Proteins  Gold compounds 
Mechanism of action  Mass spectrometry
Introduction
A few recent studies highlighted the importance of gold
compounds as a new family of cytotoxic agents with the
123
1294
potential of becoming effective anticancer drug candidates
[1–3]. Indeed, a variety of gold compounds, either gold(III)
or gold(I), bearing different structural motifs were reported
to cause effective cell death in vitro [4–7] in numerous
cancer cell lines; for a few gold(III) compounds, i.e.
gold(III) porphyrins and gold(III) dithiocarbamates, pre-
liminary but very promising in vivo results were also
obtained [1, 8].
The modes of action of antiproliferative gold com-
pounds are still largely unknown and a matter of intense
research and debate. They appear to be multifaceted and
deeply distinct from those of clinically established, DNA-
damaging platinum compounds. Notably, there is a grow-
ing body of evidence suggesting that some selected protein
targets rather than nucleic acids primarily mediate the
biological effects of cytotoxic gold compounds [9, 10]. A
few studies postulated that the cytotoxic effects of gold
compounds are driven by a specific antimitochondrial
mechanism, ultimately leading to apoptotic cancer cell
death [11–13]; however, this latter hypothesis still requires
conclusive validation.
Over the last 30 years, a number of studies have
appeared concerning the reactions of gold compounds with
proteins, analysed at the molecular level. Some pioneering
work was conducted by Frank Shaw and co-workers [14]
into the reactions of auranofin with serum albumin. Other
important studies were reported by Peter Sadler and
Fig. 1 Structures of the various
gold compounds considered in
this study: a auranofin
(Mr 678.50); b Auoxo6
(Mr 1,084.41); c Au2phen
(Mr 1,132,39); d Aubipyc
(Mr 634,32)
b c
123
J Biol Inorg Chem (2012) 17:1293–1302
co-workers [15, 16] concerning serum albumin and cy-
clophilin. Talib et al. [17] used mass spectrometry (MS)
methods to investigate the binding of various gold(I) com-
pounds to serum albumin. A systematic study conducted in
our laboratory a few years ago considered the reactions of
various gold compounds with serum albumin and investi-
gated the underlying interaction mechanisms [9, 18, 19].
Very recent studies, mainly based on electrospray ionisa-
tion (ESI) MS, analysed the reactions of cytotoxic gold
compounds with cytochrome c and lysozyme and descri-
bed the resulting adducts [20, 21].
During the last few years, we have prepared and char-
acterised three novel gold(III) complexes—namely
[(bipy2Me)2Au2(l-O)2][PF6]2 (where bipy2Me is 6,60 -dime-
thyl-2,20 -bipyridine) (Auoxo6), [(phen2Me)2Au2
2Me
(l-O)2][PF6]2 (where phen is 2,9-dimethyl-1,10-phe-
nanthroline) (Au2phen) and [(bipydmb-H)Au(OH)][PF6]
(where bipydmb-H is deprotonated 6-(1,1-dimethylbenzyl)-
2,20 -bipyridine) (Aubipyc)—that revealed quite promising
antiproliferative properties when tested in vitro on a variety
of cancer cell lines. The three complexes mentioned are
shown in Fig. 1; the schematic structure of auranofin, a
clinically established gold(I) drug used here for compari-
son purposes, is also reported.
Auoxo6 and Au2phen are dioxo-bridged dinuclear
gold(III) complexes with 6,60 - dimethyl-2,20 -bipyridyne
and 2,9-dimethyl-1,10-phenanthroline, respectively, as
a
J Biol Inorg Chem (2012) 17:1293–1302
ancillary ligands [22–24]. Both gold(III) centres in Auoxo6
and Au2phen display a classic square-planar coordination,
with slight square-pyramidal distortions. Aubipyc is a
gold(III) cyclometalated derivative of 6-(1,1-dimethylben-
zyl)-2,20 -bipyridine, which features an N,N,C sequence of
donor atoms of the terdentate bipyridine ligand and an
oxygen atom of a hydroxo ligand coordinated to the
gold(III) centre in a square-planar geometry [25]. Auran-
ofin, 2,3,4,6-tetra-O-acetyl-1-thio-b-D-pyranosato-S-(tri-
ethylphosphine)gold(I), is a linear gold(I) complex of the
type R3P-Au(I)-SR0 , where SR0 is a sugar thiolate.
The cytotoxic properties of the above-mentioned
gold(III) complexes were evaluated in vitro on a standard
36 cancer cell line panel available in Oncotest (Freiburg,
Germany) according to established procedures [26, 27].
The COMPARE algorithm [28, 29] applied to the analysis
of the growth inhibition data obtained revealed that the
profiles of Au2phen are very similar to those of Auoxo6;
this behaviour reflects the pronounced structural analogy
between these two compounds [22]. Tentatively, the patterns
of antiproliferative activity obtained for Au2phen and Auoxo6
were referred to inhibition of histone deacetylase on the basis
of bioinformatic analysis. At variance, Aubipyc, although
being appreciably cytotoxic to A2780 ovarian cancer cells,
was only moderately effective in the Oncotest panel. Auran-
ofin was confirmed to be highly cytotoxic in the Oncotest
panel, with a mode of action resembling that of typical anti-
proteasomal agents [8]. As mentioned above, correlations in
the antiproliferative profiles between the various gold com-
pounds and cisplatin were very poor, implying the occurrence
of drastically different modes of action.
On the whole, COMPARE analysis of these compounds
as well as of several other cytotoxic gold compounds
suggested that a variety of proteins such as histone
deacetylase, cyclin-dependent kinase, proteasome proteins
and mammalian target of rapamycin might be reliable
biomolecular targets and thus account for their biological
effects; in other words, the biological actions of the gold
compounds investigated might be best interpreted in terms
of metalation and inactivation of a few crucial intracellular
proteins that are effective cancer targets.
The great current interest in interactions between metal-
based drugs and proteins prompted us to conduct a sys-
tematic analysis of the reactions occurring between the
above-mentioned gold compounds and two model proteins,
namely horse heart cytochrome c and hen egg white
lysozyme. These small proteins are very amenable to ESI
MS analysis and were specifically chosen for the present
investigation. The interactions of these two proteins with a
variety of metalbased drugs, and in particular with cis-
platin, were previously investigated in detail [30, 31].
Cytochrome c was found to produce several platinum
adducts with various metal-to-protein stoichiometries as it
1295
possesses at least four strong anchoring sites for metals; in
contrast, lysozyme manifested a lower tendency to form
platinum adducts, and typically yielded monometalated
platinum derivatives through platinum coordination to its
unique histidine residue, i.e. His-15.
The main goal of the present study concerning interac-
tions of a few gold-based drugs with proteins is to identify
and define, at the molecular level, the modes of protein
metalation by these gold compounds. Metalation is the
process through which proteins are modified upon reaction
with metal-containing agents; such reactions typically
involve formation of adducts between the protein and a
metal species (often a molecular fragment derived from the
starting metal compound). The metallic fragment may be
bound to the protein either by a direct coordinative bond or
through a non-covalent interaction. Precise information on
the actual metalation mechanisms may be achieved quite
realistically through a detailed structural characterisation of
metal–protein adducts including adduct quantification with
respect to native unmodified proteins, identification of the
nature and the number of protein-bound metallic fragments,
and localisation of the anchoring sites for metallic fragments.
ESI MS appears to be a very appropriate and powerful tool to
gain this kind of information; conversely, absorption spec-
troscopy offers the opportunity to monitor continuously, in
‘‘real time’’, the various metallodrug–protein species.
Materials and methods
Metal complexes and proteins
Au2phen [21], Auoxo6 [6] and Aubipyc [25] were syn-
thesised and characterised as described previously (see the
respective references). Auranofin was purchased from
Vinci-Biochem. Horse heart cytochrome c (C7752) and
chicken hen egg white lysozyme (L7651) were purchased
from Sigma and used as received.
Physicochemical measurements
UV–vis spectra were recorded with a Varian Cary 50
UV–vis spectrophotometer. Mass spectra were recorded
with an LTQ Orbitrap high-resolution mass spectrometer
(Thermo Scientific, San Jose, CA, USA) equipped with a
conventional ESI source.
UV–vis spectrophotometric studies
Small amounts of freshly prepared concentrated solutions
of the individual complexes in dimethyl sulfoxide (or
EtOH/H2O for auranofin) were diluted in the reference
buffer (10 mM phosphate, pH 7.4). The concentration of
123
1296
each gold compound in the final sample was 10-5 M. The
resulting solutions were monitored by collection of the
electronic spectra for 24 h at room temperature.
Interactions with proteins
Electronic spectra of the model protein (lysozyme or
cytochrome c) at 10-5 M were recorded before and after
the addition of each gold complex at a stoichiometric ratio
of 3:1 (metal to protein) for 24 h at room temperature in
10 mM phosphate buffer, pH 7.4. Similar spectrophoto-
metric studies were conducted in the presence of a reducing
agent, i.e. glutathione (GSH) or ascorbic acid, added to the
various reaction batches to a final concentration of 1 mM.
Sample preparation and MS analysis
Metal complex–protein adducts were prepared by mixing
equivalent amounts of the three proteins (10-4 M) in
25 mM tetramethylammonium acetate buffer, pH 7.4.
Then, the three gold(III) complexes and auranofin were
added (3:1 metal-to-protein ratio) to the solution and the
mixture was incubated at 37 °C for 72 h. After a 20-fold
dilution with water, ESI MS spectra were recorded by
direct introduction of the sample at a flow rate of 5 ll/min
into an Orbitrap high-resolution mass spectrometer
(Thermo Scientific, San Jose, CA, USA) equipped with a
conventional ESI source. The working conditions were as
follows: spray voltage 3.1 kV, capillary voltage 45 V and
capillary temperature 220 °C. The sheath and the auxiliary
gases were set at 17 (arbitrary units) and 1 (arbitrary unit),
respectively,. For acquisition, Xcalibur 2.0 (Thermo Sci-
entific) was used and monoisotopic and average deconvo-
luted masses were obtained by using the integrated Xtract
tool. For spectrum acquisition, a nominal resolution (at m/
z 400) of 100,000 was used.
Results
Solution behaviour of the three gold(III) compounds
Before studying their interactions with proteins, we ana-
lysed the solution behaviour of the three gold(III) com-
plexes under well-controlled experimental conditions.
UV–vis absorption spectroscopy was selected as the
method to monitor continuously the behaviour of their
gold(III) chromophores in the reference buffer (10 mM
phosphate buffer, pH 7.4). All three gold(III) compounds
manifested appreciable stability when monitored for 24 h
at 20 °C, as documented by the substantial invariance of
their absorption spectra (see Fig. S1). These observations
are in agreement with previous spectrophotometric results
123
J Biol Inorg Chem (2012) 17:1293–1302
Fig. 2 Time course UV–vis spectra of Auoxo6 dissolved in 10 mM
phosphate buffer, pH 7.4, before and after addition of 1 mM
glutathione. The concentration of the complex was 10-5M. The
spectra were recorded before (red line) and after the addition of the
reducing agent at 0 min (blue line), 5 min (green line), 10 min (violet
line), 15 min (grey line) and 1 h (light-blue line)
obtained for the same compounds under slightly different
experimental conditions [6, 25]. It is remarkable that the
?3 oxidation state is essentially conserved in all cases over
the whole observation period.
In contrast, important spectral changes were highlighted
when these gold(III) complexes were challenged with an
excess of two biologically occurring reducing agents, i.e.
ascorbic acid and GSH (Fig. S2). Auoxo6 and Au2phen
underwent quick and facile reduction upon addition of an
excess of either reducing agent, which was documented by
evident spectral changes. The reaction of Auoxo6 with
GSH is very representative as shown in Fig. 2; indeed, the
ligand-to-metal charge transfer band centred at 325 nm
rapidly disappears upon addition of GSH. Conversely, in
the case of Au2phen, we noticed large changes in the shape
and intensity of the composite band centred at 280 nm that
are suggestive of reduction of gold(III) (Fig. S2). Aubipyc
was far more resistant to reduction: indeed, ascorbic acid
failed to reduce it, whereas an excess of GSH induced a
progressive–although slow—reduction of the gold(III)
centre (Fig. S2).
Reactions of gold compounds with model proteins:
spectrophotometric analysis
The interactions of these gold(III) compounds with the two
aforementioned model proteins were explored. Each of the
gold(III) compounds was reacted with cytochrome c or
lysozyme at a standard metal-to-protein molar ratio of 3:1
according to previously defined experimental procedures
and the resulting solutions were incubated for 72 h at
37 °C (25 mM tetramethylammonium acetate buffer, pH
7.4). Samples were continuously monitored through spec-
trophotometric analysis. For comparison purposes, similar
experiments were conducted using auranofin as the
J Biol Inorg Chem (2012) 17:1293–1302 1297

metalating agent. At the end of the incubation period,
samples were analysed by ESI MS as described in the next
section.
UV–vis spectrophotometric analysis of metallodrug–
protein samples allows the continuous monitoring of the
various gold(III) centres in the presence of these model
proteins. In the case of cytochrome c—i.e. a metalloprotein
characterised by intense transitions in the visible region—it
is possible to monitor simultaneously both the metallo-
drug’s chromophore and the active site of the protein in the
course of their interaction.
From spectral inspection, it is apparent that addition of
protein does not affect the behaviour of the various
gold(III) chromophores (Fig. 3). Yet, modest—but slowly
progressive—spectral changes were detected in the case of
the reaction of Auoxo6 with cytochrome c, as shown in
Fig. 3. Some significant changes in the shape of the com-
posite band of Au2phen centred at 280 nm were also
detected that are suggestive of the occurrence of partial
reduction of gold(III). Conversely, the spectra of the vari-
ous metallodrug–cytochrome c systems reveal that the
protein chromophore is substantially stable for 24 h, with
cytochrome c remaining in its oxidised ferric form.
ESI MS spectra of metallodrug–protein samples
ESI MS is a very powerful analytical tool to characterise
metallodrug–protein interactions at the molecular level [32,
33]. Accordingly, ESI MS analysis of the previously
described samples allowed us to identify and characterise a
number of metal–protein adducts whose presence is hardly
documented in the electronic absorption spectra owing to
the intrinsic limitations of that technique. In particular, ESI
MS measurements permitted the nature of protein-bound
metallic fragments to be disclosed and their binding stoi-
chiometries to be determined, thus providing indirect
mechanistic insight into the respective metalation pro-
cesses. As a variety of different situations were encoun-
tered depending on the nature of the metallodrug and, to a
lesser extent, of the protein, the various cases will be
illustrated separately.
Auoxo6 and Au2phen
These two gold complexes are described together as they
manifested rather similar reactivity patterns with regard to
both lysozyme and cytochrome c. ESI MS spectra of

Fig. 3 Time course UV–vis

spectra of gold compounds
dissolved in 10 mM phosphate
buffer, pH 7.4, upon 24 h
incubation with lysozyme or
cytochrome c. The protein
concentration was 10-5 M and
the metallodrug-to-protein
molar ratio was 3:1. The plots
on the left show the spectral
features of cytochrome c upon
addition of Au2phen (a),
Auoxo6 (b) and Aubipyc (c).
The plots on the right show the
spectral features of lysozyme
upon addition of Au2phen (d),
Auoxo6 (e) and Aubipyc (f). All
reported spectra were recorded
before (red line) and after the
addition of the three gold
compounds at 0 h (green line),
1 h (violet line), 12 h (blue line)
and 24 h (grey line)

123
1298
Fig. 4 LTQ Orbitrap electrospray ionisation (ESI) mass spectra of
gold compounds dissolved in 25 mM tetramethylammonium acetate
(TMeAmAc) buffer, pH 7.4, in the presence of the two representative
proteins (lysozyme and cytochrome c) after 72 h incubation at 37 °C.
The protein concentration was 10-4 M (with a metal complex to
protein molar ratio of 3:1). The plots on the left show Au2phen in the
protein solutions incubated with these two compounds for
72 h at a 3:1 metallodrug-to-protein molar ratio are shown
comparatively in Fig. 4.
From careful inspection of this Fig. 4, it is evident that a
number of metal–protein adducts are formed in the various
cases, although in greatly different amounts, as can be
judged by comparing the relative peak intensities. Au2phen
manifests a very high reactivity with cytochrome c, which
is accompanied by the formation of relatively large
amounts of adducts; in the other cases, only small amounts
of metal–protein adducts are formed. In all cases, the
123
J Biol Inorg Chem (2012) 17:1293–1302
presence of cytochrome c (A) or lysozyme (B). The plots on the right
show Auoxo6 in the presence of cytochrome c (C) or lysozyme (D).
All spectra show the peak of the native protein (a) and of gold(I)–
protein adducts with stoichiometries of 1:1 (b), 2:1 (c), 3:1 (d), 4:1
(e) and 5:1 (f)
position of the peaks permits us to assign these adducts to
one or more ‘‘naked’’ gold(I) ions binding to protein; no
evidence of ‘‘mass shifts’’ corresponding to the original
gold(III) ligands was obtained, implying that reduction of
gold(III) and complex disruption always precede protein
binding. Very interestingly, cytochrome c favours the for-
mation of a metal protein adduct with a stoichiometry of
4:1, i.e. four gold(I) ions bound to the protein (see the peak
with molecular mass of 13,142 Da); this observation is in
agreement with previous studies on the reaction of cyto-
chrome c with gold saccharinate compounds [20] and with
J Biol Inorg Chem (2012) 17:1293–1302
Fig. 5 LTQ Orbitrap ESI mass spectra of Aubipyc dissolved in
25 mM TMeAmAc buffer, pH 7.4, in the presence of lysozyme
(A) and cytochrome c (B) after 72 h incubation at 37 °C. The protein
concentration was 10-4 M (with a metal complex to protein molar
the presence of four major anchoring sites for metals on the
cytochrome c surface, i.e. Met-65, Met-80, His-18 and
His-33.
Aubipyc
When challenged with cytochrome c and lysozyme,
Aubipyc manifested a quite constant behaviour that is
different from that of Auoxo6 and Au2phen. Small amounts
of metal–protein adducts were formed with both proteins;
in both cases the protein-bound molecular fragment was
the same, corresponding to the [(bipydmb-H)Au] moiety,
i.e. the gold(III) centre plus the N,N,C terdentate ligand, as
proved by ESI MS analysis (Fig. 5).
That gold in the Aubipyc–protein adducts remains in the
?3 oxidation state is in good agreement with previous
results documenting the high redox stability of this or-
ganogold(III) compound and its relatively scarce reactivity
[25].
Auranofin
The reaction profiles of auranofin with the two model
proteins were also investigated for comparative purposes.
ESI MS spectra obtained at 72 h for auranofin–protein
adducts are shown in Fig. 6. The ESI MS spectra document
well the formation of metallodrug–protein adducts,
although in modest amounts. In both cases, the position of
the major peak for the metal–protein adducts is consistent
with the intact auranofin molecule binding to protein,
1299
ratio of 3:1). Both spectra show peaks corresponding to the native
protein (a) and the [(bipydmb-H)Au] moiety–protein adduct [where
bipydmb-H is deprotonated 6-(1,1-dimethylbenzyl)-2,20 -bipyridine; Mr
472.36] (b)
probably by non-covalent interaction. The amount of
adduct formed is nearly the same. Also, there is some
evidence in both cases for the formation of additional
adducts where the [(triethylphosphine)gold(I)] moiety is
protein-bound.
Reactions of gold compounds with model proteins
in the presence of reducing agents
The results reported in the previous sections suggest that
reduction of gold(III) may play a crucial role in the protein
metalation process, especially in the case of Auoxo6 and
Au2phen. This observation led us to explore the effects of
the addition of an excess of ‘‘physiological’’ reducing
agents, such as ascorbic acid and GSH, to the various
reaction batches on the assumption that a strongly reducing
environment might favour the protein metalation process.
Indeed, it is well known that hypoxic cancer tissues exhibit
a frankly reducing microenviroment. Studies were restric-
ted to Auoxo6 and Au2phen because the gold(III) centre in
Aubipyc is known to be highly resistant to reduction (see
earlier).
Auoxo6 and Au2phen were dissolved in the buffer in the
presence of either cytochrome c or lysozyme at standard
3:1 molar ratios and ascorbic acid or GSH was added. The
concentration of the reducing agent was fixed at 1 mM as
this value reflects the typical intracellular concentrations of
GSH. Remarkably, both reducing agents caused the char-
acteristic ligand-to-metal charge transfer bands of Auoxo6
to disappear quickly, implying rapid reduction of the two
123
1300
Fig. 6 LTQ Orbitrap ESI mass spectra of auranofin dissolved in
25 mM TMeAmAc buffer, pH 7.4, in the presence of cyto-
chrome c (A) or lysozyme (B) after 72 h incubation at 37 °C. The
protein concentration was 10-4 M (with a metal complex to protein
gold(III) centres, and cytochrome c experienced simulta-
neous conversion to its reduced form; important spectral
changes were also observed for Au2phen in line with those
previously described for the reduction of Au2phen alone in
the reference buffer (Fig. S3).
The resulting metallodrug–protein samples (Auoxo6 and
Au2phen reacted with cytochrome c or lysozyme) were
then analysed by ESI MS after short incubation times.
Comparative analysis of the spectra allows us to state that
addition of ascorbic acid invariantly causes the formation
of far greater amounts of metal–protein adducts. At vari-
ance, GSH, although accelerating the reduction of the
gold(III) centres, apparently does not induce greater adduct
formation. This may be explained by assuming that GSH at
high concentrations (1 mM, as is the case here) is capable
of sequestering most gold(I) ions in the form of Au(GSH)2
complexes; yet, the presence of small quantities of adducts
in which gold(I) ions are attached to these two proteins
could be demonstrated (Fig. S4).
Discussion
Gold compounds form a new class of promising cytotoxic
metallodrugs that are attracting growing attention within
the scientific community as potential anticancer agents.
The modes of action of cytotoxic gold compounds are still
poorly understood. However, a number of observations
suggest that cytotoxic gold compounds, at variance with
123
J Biol Inorg Chem (2012) 17:1293–1302
molar ratio of 3:1. Both spectra show peaks corresponding to the
native protein (a) and to the formation of a [(triethylphos-
phine)gold(I)] moiety–protein adduct (Mr 315) (b) and an aurano-
fin–protein adduct with stoichiometries of 1:1 (c) and 2:1 (d)
established platinum(II) drugs, are scarcely reactive
towards DNA. Accordingly, their biological actions are
most probably the consequence of tight interactions
occurring with a number of intracellular protein targets.
Inactivation of some crucial proteins might indeed consti-
tute the ‘‘decisive event’’ ultimately triggering apoptotic
cell death. In particular, the selenoenzyme thioredoxin
reductase seems to be a good candidate protein target for
several cytotoxic gold compounds. A variety of other
protein targets have been proposed [34–36], but conclusive
target validation is still lacking.
In this work, we have analysed—systematically—the
interactions of three representative gold(III) compounds
with two model proteins—i.e. lysozyme and cyto-
chrome c—to reveal the molecular details of the inherent
metalation processes. Comparative studies were conducted
on auranofin, a clinically established antiarthritic gold(I)
drug. All these gold compounds typically behave as classic
prodrugs; upon ‘‘chemical activation’’, they react with
proteins and form stable adducts. Complex activation may
simply consist of a ligand replacement reaction but also of
reductive disruption of the gold(III) centre and consequent
formation of ‘‘naked’’ gold(I) ions.
The metalation processes of the model proteins men-
tioned were investigated—independently—by absorption
spectroscopy and ESI MS methods. Absorption spectros-
copy allowed us to monitor continuously the various
metallodrug–protein systems over the whole incubation
time of 72 h. Conversely, ESI MS spectra recorded on
J Biol Inorg Chem (2012) 17:1293–1302
metallodrug–protein samples at the end of the incubation
period were particularly informative in revealing adduct
formation and in determining the final metal-to-protein
stoichiometry and the nature of protein-bound metallic
fragments.
Protein metalation could be documented in most cases,
on the basis of the identification of a variety of metal–
protein adducts in the ESI MS spectra. The number and
the nature of the protein-bound metallic species were
determined with accuracy. A rough estimate of the
amount of protein metalation was achieved by comparing
the experimental peak intensities—i.e. the peak intensity
of the free protein versus the peak intensities of its metal
adducts.
Three distinct kinds of metallodrug–protein adducts
emerged from our studies. These adducts are formed
through one of the following mechanisms: (1) reduction of
gold(III), complex disruption and coordinative binding of
the resulting gold(I) ions to a typical protein donor, e.g.
histidine or methionine; (2) coordinative binding of a
redox-stable gold(III) fragment to protein donors; (3) non-
coordinative binding of the intact gold complex to the
protein. The pathway actually followed for adduct forma-
tion will typically depend on the nature of the metal
complex and the nature of the interacting protein.
Large differences were highlighted in the relative effi-
ciency of the various metalation processes and, accord-
ingly, in the quantities of adducts formed. Differences in
protein metalation processes may be very relevant for the
effective mechanism of action of cytotoxic gold(III) com-
pounds. Notably, the degree of protein metalation and the
resulting metal–protein stoichiometry critically depend on
both the nature of the protein and the nature of the metal
complex. The cases of the reactions of cytochrome c and
lysozyme with Au2phen and Auoxo6 are particularly
instructive; indeed, owing to the availability of a greater
number of metal-binding side chains, cytochrome c gives
rise to adducts of higher metal-to-protein stoichiometry
than lysozyme. Particularly interesting is the reaction of
Au2phen with cytochrome c that leads to the formation of
large amounts of a tetragold cytochrome c adduct where
four gold(I) ions are tightly associated with the protein.
Further studies are under way to better characterise this
adduct.
Typically, our studies were conducted according to a
standard protocol with a long incubation time of 72 h.
However, some additional experiments were performed on
cytochrome c with shorter incubation times (see Figs. S5,
S6). Reduction of the incubation time resulted in a drastic
reduction in the amounts of adducts formed, with no evi-
dent change in the nature of the adducts.
Overall, the present investigation stresses the biological
and pharmacological relevance of protein metalation
1301
processes that seem to play a central role in determining
the cellular effects of cytotoxic gold compounds (and more
in general of various families of metal-based drugs). It is
reasonable to assume that the cytotoxic gold(III) com-
plexes that were selected for the present investigation have
the potential of metalating and altering a great number of
cellular proteins according to the mechanisms described
herein; in turn, selective inactivation of a few crucial
proteins (the true pharmacological targets) may ultimately
cause irreversible cell damage and death. A lot of experi-
mental effort is still needed to elucidate these issues in
depth, this being one of the major missions of modern
metallomics.
The role of reducing agents in the protein metalation
processes was also assessed through a few additional
experiments. Although it is evident that protein metalation
in two specific cases requires reduction of gold(III) to
gold(I), we established that the presence of an excess of
two distinct biological reducing agents, i.e. GSH and
ascorbic acid, while causing rapid reduction of the gold(III)
centres, and also of cytochrome c, had contrasting effects
on the formation of metal–protein adducts. Ascorbic acid
greatly favoured adduct formation in the reactions of Au-
oxo6 and Au2phen with cytochrome c or lysozyme; in
contrast, GSH did not exhibit a similar effect, possibly
because of extensive complexation of gold(I) by GSH
itself.
In conclusion, with the present study, we have tried to
elucidate, at the molecular level, the processes of protein
metalation that occur when a few representative cytotoxic
gold compounds react with two model proteins. A variety
of situations were encountered depending on the nature of
the metal complex and the nature of the protein. From the
results obtained, we can state that the process of adduct
formation may follow a number of distinct routes leading
to a variety of structurally different adducts. However,
prediction of which route will be preferred and what kinds
and amounts of metal–protein adducts will be formed for a
metallodrug–protein pair remains a very hard task; in fact,
the protein metalation processes appear to depend greatly
both on the nature of the metal complex and on the nature
of the protein.
It would be reasonable to envision that in defining the
in vivo pathways of metal-based drugs an important role is
played by proteins involved in the homeostasis of transition
metal ions such as copper, zinc and nickel and assembly of
their metal cofactors in metalloenzymes [37–40] but this
aspect has been little characterised until now.
We also showed that creating a strongly reducing
microenvironment resulted in variable effects on the pro-
tein metalation processes depending on the nature of the
reducing agent applied. It follows that each case of interest
is yet to be investigated individually and that specific
123
1302 J Biol Inorg Chem (2012) 17:1293–1302

trends will hopefully be defined when a far larger mass of 20. Maiore L, Cinellu MA, Michelucci E, Moneti G, Nobili S,
experimental data are available. Landini I, Mini E, Guerri A, Gabbiani C, Messori L (2011) J
Inorg Biochem 105:348–355
21. Gabbiani C, Casini A, Kelter G, Cocco F, Cinellu MA, Fiebig
Acknowledgments We gratefully acknowledge Beneficentia Stif- HH, Messori L (2011) Metallomics 3:1318–1323
tung (Vaduz, Liechtenstein) and AIRC (IG-12085) for generous 22. Cinellu MA, Maiore L, Manassero M, Casini A, Arca M, Fiebig
financial support. HH, Kelter G, Michelucci E, Pieraccini G, Gabbiani C, Messori L
(2010) ACS Med Chem Lett 1:336–339
23. Gabbiani C, Casini A, Messori L, Guerri A, Cinellu MA, Min-
ghetti G, Corsini M, Rosani C, Zanello P, Arca M (2008) Inorg
References Chem 47:2368–2379
24. Cinellu MA, Minghetti G, Pinna MV, Stoccoro S, Zucca A,
Manassero M, Sansoni M (1998) J Chem Soc Dalton Trans
1. Nobili S, Mini E, Landini I, Gabbiani C, Casini A, Messori L
1735–1741
(2010) Med Res Rev 30:550–580
25. Marcon G, Carotti S, Coronnello M, Messori L, Mini E, Orioli P,
2. Gabbiani C, Casini A, Messori L (2007) Gold Bull 40:73–81
Mazzei T, Cinellu MA, Minghetti G (2002) J Med Chem
3. Tiekink ERT (2002) Crit Rev Oncol Hematol 42:225–248
45:1672–1677
4. Vergara E, Casini A, Sorrentino F, Zava O, Cerrada E, Rigobello
26. Casini A, Kelter G, Gabbiani C, Cinellu MA, Minghetti G,
MP, Bindoli A, Laguna M, Dyson PJ (2010) Chem Med Chem
Fregona DS, Fiebig HH, Messori L (2009) J Biol Inorg Chem
5:96–102
14:1139–1149
5. McKeage MJ, Maharaj L, Berners-Price SJ (2002) Coord Chem
27. http://www.oncotest.de/. Accessed 31 Oct 2012
Rev 232:127–135
28. Paull KD, Shoemaker RH, Hodes L, Monks A, Scudiero DA,
6. Casini A, Cinellu MA, Minghetti G, Gabbiani C, Coronnello M,
Rubinstein L, Plowman J, Boyd MR (1989) J Natl Cancer
Mini E, Messori L (2006) J Med Chem 49:5524–5531
81:1088–1092
7. Che C-M, Sun RW-Y, Yu WY, Ko CB, Zhu NY, Sun HZ (2003)
29. Huang RL, Wallqvist A, Covell DG (2005) Biochem Pharmacol
Chem Commun 1718–1719
69:1009–1039
8. Milacic V, Chen D, Ronconi L, Landis-Piwowar KR, Fregona D,
30. Zhang N, Du Y, Cui M, Xing J, Liu Z, Liu S (2012) Anal Chem
Dou QP (2006) Cancer Res 66:10478–10486
84(14):6206–6212
9. Gabbiani C, Messori L (2011) Anti Cancer Ageing Med Chem
31. Tanley SW, Schreurs AM, Kroon-Batenburg LM, Meredith J,
11:929–939
Prendergast R, Walsh D, Bryant P, Levy C, Helliwell JR (2012)
10. Mirabelli CK, Sung CM, Zimmerman JP, Hill DT, Mong S,
Acta Crystallogr D Biol Crystallogr 68(5):601–612
Crooke ST (1986) Biochem Pharmacol 35:1427–1433
32. Timerbaev AR, Pawlak K, Gabbiani C, Messori L (2011) Trends
11. Bindoli A, Rigobello MP, Scutari G, Gabbiani C, Casini A,
Anal Chem 30:1120–1138
Messori L (2009) Coord Chem Rev 253:1692–1707
33. Casini A, Guerri A, Gabbiani C, Messori L (2008) J Inorg Bio-
12. Rubbiani R, Kitanovic I, Alborzinia H, Can S, Kitanovic A,
chem 102:995–1006
Onambele LA, Stefanopoulou M, Geldmacher Y, Sheldrick WS,
34. Arnér ES, Holmgren A (2006) Semin Cancer Biol 6:420–426
Wolber G, Prokop A, Wölfl S, Ott I (2010) J Med Chem
35. Nguyen P, Awwad RT, Smart DD, Spitz DR, Gius D (2006)
53:8608–8618
Cancer Lett 236:164–174
13. Hickey JL, Ruhayel RA, Barnard PJ, Baker MV, Berners-Price
36. Vergara E, Casini A, Sorrentino F, Zava O, Cerrada E, Rigobello
SJ, Filipovska A (2008) J Am Chem Soc 130:12570–12571
MP, Bindoli A, Laguna M, Dyson PJ (2010) Chem Med Chem
14. Roberts JR, Xiao J, Schliesman B, Parsons DJ, Shaw CF III
5(1):96–102
(1996) Inorg Chem 35:424–433
37. Banci L, Bertini I, Cantini I, Ciofi-Baffoni S (2010) Cell Mol Life
15. Christodoulou J, Sadler PJ, Tucker A (1994) Eur J Biochem
Sci 67(15):2563–2589
225:363–368
38. Costello LC, Fenselau CC, Franklin RB (2011) J Inorg Biochem
16. Zou J, Taylor P, Dornan J, Robinson SP, Walkinshaw MD, Sadler
105(5):589–599
PJ (2000) Angew Chem Int Ed 39:2931
39. Zambelli B, Musiani F, Benini S, Ciurli S (2011) Acc Chem Res
17. Talib J, Beck JL, Ralph SF (2006) J Biol Inorg Chem 11:559–570
44(7):520–530
18. Marcon G, Messori L, Orioli P, Cinellu MA, Minghetti G (2003)
40. Zambelli B, Cremades N, Neyroz P, Turano P, Uversky V, Ciurli
Eur J Biochem 270:4655–4661
S (2012) Mol Biosyst 8:220–228
19. Messori L, Balerna A, Ascone I, Castellano C, Gabbiani C, Ca-
sini A, Marchioni C, Jaouen G, Congiu Castellano A (2011) J
Biol Inorg Chem 16:491–499

123

View publication stats