A study of the glutathione metaboloma peptides by energy-resolved mass spectrometry as a tool to investigate into the interference of toxic heavy metals with their metabolic processes

Article abstract of DOI:10.1002/jms.1143

To better understand the fragmentation processes of the metal-biothiol conjugates and their possible significance in biological terms, an energy-resolved mass spectrometric study of the glutathione conjugates of heavy metals, of several thiols and disulfides of the glutathione metaboloma has been carried out. The main fragmentation process of γ-glutamyl compounds, whether in the thiol, disulfide, thioether or metal-bis-thiolate form, is the loss of the γ-glutamyl residue, a process which ERMS data showed to be hardly influenced by the sulfur substitution. However, loss of the γ-glutamyl residue from the mono-S-glutathionyl-mercury (II) cation is a much more energetic process, possibly pointing at a strong coordination of the carboxylic group to the metal. Moreover, loss of neutral mercury from ions containing the γ-glutamyl residue to yield a sulfenium cation was a much more energetic process than those not containing them, suggesting that the redox potential of the thiol/disulfide system plays a role in the formal reduction of the mercury dication in the gas phase. Occurrence of complementary sulfenium and protonated thiol fragments in the spectra of protonated disulfides of the glutathione metaboloma mirrors the thiol/disulfide redox process of biological importance. The intensity ratio of the fragments is proportional to the reduction potential in solution of the corresponding redox pairs. This finding has allowed the calculation of the previously unreported reduction potentials for the disulfide/thiol pair of cysteinylglycine, thereby confirming the decomposition scheme of bis- and mono-S-glutathionyl-mercury (II) ions. Finally, on the sole basis of the mass spectrometric fragmentation of the glutathione-mercury conjugates, and supported by independent literature evidence, an unprecedented mechanism for mercury ion-induced cellular oxidative stress could be proposed, based on the depletion of the glutathione pool by a catalytic mechanism acting on the metal (II)-thiol conjugates and involving as a necessary step the enzymatic removal of the glutamic acid residue to yield a mercury (II)-cysteinyl-glycine conjugate capable of regenerating neutral mercury through the oxidation of glutathione thiols to the corresponding disulfides. Copyright

Full text of DOI:10.1002/jms.1143

JOURNAL OF MASS SPECTROMETRY
J. Mass Spectrom. 2006; 41: 1578–1593
Published online in Wiley InterScience
(www.interscience.wiley.com) DOI: 10.1002/jms.1143
A study of the glutathione metaboloma peptides by
energy-resolved mass spectrometry as a tool to
investigate into the interference of toxic heavy metals
with their metabolic processes†
Federico Maria Rubino,1∗ Marco Pitton,2 Gabri Brambilla2 and Antonio Colombi2
1
Laboratory for Analytical Toxicology and Metabolomics, Departments of Medicine, Surgery and Odontology, Universita` degli Studi di Milano at
´
Ospedale San Paolo, v. A. di Rudinı 8 I-20142 Milano, Italy
2
´
Occupational Medicine, Universita` degli Studi di Milano at Ospedale San Paolo, v. A. di Rudinı 8 I-20142 Milano, Italy
Received 31 May 2006; Accepted 16 October 2006
To better understand the fragmentation processes of the metal–biothiol conjugates and their possible
significance in biological terms, an energy-resolved mass spectrometric study of the glutathione conjugates
of heavy metals, of several thiols and disulfides of the glutathione metaboloma has been carried out.
The main fragmentation process of g-glutamyl compounds, whether in the thiol, disulfide, thioether
or metal-bis-thiolate form, is the loss of the g-glutamyl residue, a process which ERMS data showed
to be hardly influenced by the sulfur substitution. However, loss of the g-glutamyl residue from the
mono-S-glutathionyl-mercury (II) cation is a much more energetic process, possibly pointing at a strong
coordination of the carboxylic group to the metal. Moreover, loss of neutral mercury from ions containing
the g-glutamyl residue to yield a sulfenium cation was a much more energetic process than those not
containing them, suggesting that the redox potential of the thiol/disulfide system plays a role in the
formal reduction of the mercury dication in the gas phase. Occurrence of complementary sulfenium
and protonated thiol fragments in the spectra of protonated disulfides of the glutathione metaboloma
mirrors the thiol/disulfide redox process of biological importance. The intensity ratio of the fragments
is proportional to the reduction potential in solution of the corresponding redox pairs. This finding has
allowed the calculation of the previously unreported reduction potentials for the disulfide/thiol pair of
cysteinylglycine, thereby confirming the decomposition scheme of bis- and mono-S-glutathionyl-mercury
(II) ions. Finally, on the sole basis of the mass spectrometric fragmentation of the glutathione–mercury
conjugates, and supported by independent literature evidence, an unprecedented mechanism for mercury
ion-induced cellular oxidative stress could be proposed, based on the depletion of the glutathione pool
by a catalytic mechanism acting on the metal (II)–thiol conjugates and involving as a necessary step
the enzymatic removal of the glutamic acid residue to yield a mercury (II)–cysteinyl-glycine conjugate
capable of regenerating neutral mercury through the oxidation of glutathione thiols to the corresponding
disulfides. Copyright  2006 John Wiley & Sons, Ltd.
KEYWORDS: N-acetyl-cysteine; N-acetyl-penicillamine; cysteine; cysteinylglycine; electrochemical potential;
ꢀ-glutamylcysteine; glutathione; homocysteine; penicillamine; mercury; oxidative stress
as the glutathione (GSH) metaboloma; for some of the
structures as well as the biochemical relationships between
thiols, disulfides and metal (II) mercaptide conjugates,
see Scheme 1) act within the cell, both as the major
endogenous antioxidant and redox buffer system, as an
important regulator of growth and differentiation processes
and as scavengers of endogenous and xenobiotic electrophilic
metabolites and heavy metals. Supply of reduced GSH
or of suitable biosynthetic precursors from neighbouring
cells to those, such as neurons, which are unable to de
novo synthesize it, is a complex process that may be
interfered at several different levels by toxic agents, thereby
causing impairment to cellular functions caused by oxidative
INTRODUCTION
Glutathione and other metabolically related intracellular
low-molecular-weight (LMW) thiol compounds (referred to,
as a group of structural and functional cognates including
some drug agents such as N-acetyl-cysteine (NACySH),
ŁCorrespondence to: Federico Maria Rubino, Laboratory for
Analytical Toxicology and Metabolomics, Departments of
Medicine, Surgery and Odontology, Universita` degli Studi di
´
Milano at Ospedale San Paolo, v. A. di Rudinı 8 I-20142 Milano,
Italy. E-mail: federico.rubino@unimi.it
†This work was presented in part as Poster communications TuP11
and TuP21 at the 24th Informal Meeting on Mass Spectrometry
(Ustron, PL, May 13–18, 2006).
Copyright  2006 John Wiley & Sons, Ltd.

ERMS of glutathione metaboloma metal conjugates
1579
Scheme 1. Structures and partial metabolic relationships of the glutathione metaboloma thiols, disulfides and metal (II) conjugates.
stress. In fact, redox-sensitive trans-acting factors regulate
protein synthesis and degradation, signal transduction,
apoptosis and gene expression1–3 through mechanisms
mediated at the molecular level by protein glutathiolation,4–6
a post-translational modification involving formation of
mixed disulfides between GSH and key disulfide-linked
Cys residues in the native protein structure or between
glutathione disulfide (GSSG) and free protein thiol groups,
such as those observed in the ˇ-93 cysteine (CySH) of human
hemoglobin7 or in the Cys-72 of human thioredoxin.8 Similar
post-translational modifications have also been observed
with other LMW thiols, such as CySH, NACySH and
homocysteine.9–11
That several heavy metals are toxic in the human was
recognized as early as the earliest times of mining.12–14
The mechanism of the toxic activity of heavy metals is in
general terms ascribed to ‘interaction with essential thiol
groups of enzymes and proteins’, although the observed
effects on health and behavior are often specific to overexpo-
sure to specific toxic metals. As an example, mercury, lead,
cadmium and manganese all cause damages to the central
nervous system of very different nature. While manganese is
known to cause a neurologic impairment resembling either
parkinsonism15 or epilepsy,16 and cadmium a loss of olfaction
due to selective cytotoxicity of olfactory neurons,17 mercury
and lead generate well-defined and different neurobehav-
ioral syndromes, one being depressive (mercurial heretism,
testified by Lewis Carrol’s Mad Hatter, an example of an
occupational disease, frequent still down to the 1950s, and
caused in workers involved in hair-felting with mercury salt
solutions18,19), and the other excitative (according to record-
ings, an Italian worker heavily intoxicated in a tetraethyllead
manufacturing plant expired standing on his deathbed and
singing an opera aria of ‘La Traviata’20).
Structural and analytical mass spectrometry have played
a pivotal role in the characterization and measurement of
biologically occurring compounds of the GSH metaboloma,
spanning from the conjugates of electrophilic compounds
of endogenous or xenobiotic origin21,22 to the different
thiol and disulfide forms of soluble thiols,23–25 to their
disulfide conjugates with the essential or regulatory CySH
and cystine motifs which make the thiolated forms of
proteins.7–9 Conjugates of mercury and cadmium with thiols
such as GSH and CySH were identified in animals26–30 and
humans,31,32 while similar compounds with phytochelatins
were identified in metal-exposed plants.33–36 To identify such
LMW disulfides in biological samples and to investigate their
biological role, the mass spectrometric behavior of some thiol
and disulfide amino acids and peptides has been studied to
characterize the main fragmentation pathways24,25,37–40 and
to highlight the similarities and differences of the occurring
decomposition processes to those of other conjugates,
especially of the thiolates of heavy metals.41,42 With this
knowledge it is possible, in particular, to shed light on
the specific interference that individual thiol-seeking heavy
metals may display on the biological pathways in which
thiols and disulfides are involved.
Mass spectrometry is a powerful tool to this regard, since
it allows the study of the essential reactivity of biomolecules
on the basis of the unimolecular decomposition processes
occurring from ionized molecules, energized by a variety
of activation techniques. In particular, energy-resolved mass
spectrometry (ERMS) studies on triple-quadrupole tandem
mass spectrometers allow measurement of the energy
involved in any of the decomposition processes, in order
to gain insight into the involved mechanisms, which in
turn may parallel or mimic those occurring in solution. This
approach has been recently employed in the biomolecular
Copyright  2006 John Wiley & Sons, Ltd.
J. Mass Spectrom. 2006; 41: 1578–1593
DOI: 10.1002/jms

1580
F. M. Rubino et al.
field to study the reactivity of amino acids, small peptides
and nucleotides.43,44
100 mmu). To obtain fragmentation of the molecular ion in
MS/MS experiments, collisionally activated decomposition
(CAD) was accomplished with N2 as the target gas at
a pressure setting value (CAD GAS) of 1. This setting
corresponds to a reading of the ion gauge located near the
vacuum manifold of 3.0 ð 10ꢀ5 torr; the number gas density,
The mass spectrometric fragmentation of protonated di-
and tripeptides of the GSH metaboloma of the corresponding
disulfides and conjugates yields closely similar products to
those generated by enzymatic degradation in the course
of natural catabolism or through chemical and biochemical
oxidation–reduction processes.45 A deeper understanding
of the processes leading to these fragments is likely to shed
brighter light on biologically important phenomena related
to the redox behavior of individual thiol- and disulfide-
containing amino acids and peptides of the GSH metaboloma
and of their metal conjugates, as will be presented.
11
°
calculated for a manifold temperature of 310 K, is 9.3 ð 10
molecules/cm3. For comparison purposes, the pressure
reading in the absence of collision gas (CAD GAS 0) is
1.8 ð 10ꢀ5 torr, corresponding to 5.6 ð 1011 molecules/cm3,
a value comparable with other experimental conditions
reported in the literature.46 Fragment spectra were acquired
at several values of nominal collision energy ranging from
1 to 51 V in steps of 1 V increment. Acquisition of fragment
ions was from m/z 50 to a value 3–5 mass units higher than
that of the precursor ion.
EXPERIMENTAL
Reagents
Post-acquisition elaboration of the spectra
All reagents were of analytical or pharmaceutical purity
and were employed as received. The GSH metaboloma
compounds: GSH, S-methyl-glutathione (GS-Me), glu-
tathione diethyl ester (GS(Et)2), GSSG, ꢀ-glutamylcysteine
(ꢀ-GluCySH), cysteinylglycine (HSCyGly), CySH, cysteine
methylester (Cy(OMe)SH), homocysteine (hCySH), peni-
cillamine (PenSH), NACySH and N-acetyl-penicillamine
(NAPenSH) were obtained from Sigma (Prodotti Gianni,
Milano, Italy). The corresponding homodimeric disulfides,
except GSSG, were obtained by spontaneous air oxidation of
the corresponding thiol amino acids and peptides in approxi-
Post-acquisition elaboration of the spectra and calculations
were performed with custom Microsoft Excel spreadsheets,
on raw-data mass-intensity files exported from the mass
spectrometer data system as .txt files. All normalized
fragment spectra reported in the figures are plotted as Excel
(XY) dispersion graphs and not from the mass spectrometer’s
data system. In all further elaborations, the value of collision
energy was converted from the nominal, or ‘laboratory
frame’ scale (Elab) to the center-of-mass scale (Ecm) of the
impinging ion-neutral target system, calculated according to
Eqn (1),47 the mass of the target gas (mTAR) being 28, that of
nitrogen:
°
mately 10 mM water solutions at 40 C for 24 h. Heterodimeric
disulfides, such as cysteine-glutathione heterodisulfide (CyS-
SG), were prepared by reacting 10 mM GSSG with a 10-fold
Ecm D Elab ð ꢁmTAR/ꢁmTAR C mionꢂ
ꢁ1ꢂ
°
excess of the thiol (in the case of CySH) for 36 h at 40 C. The
thiol conjugates with mercury, cadmium, lead and zinc were
prepared by reacting a solution of the metal nitrate with
two equivalents of approximately 10 mM thiol solution for
For each examined compound, the measured set of
fragment spectra is composed of approximately 50 spectra,
acquired at several different values of the nominal collision
energy (defined as the potential difference between the
entrance voltage and that of the collision cell: Q0 and R02,
respectively, in the employed instrument) ranging from 0
to approximately 50 eV. Since the nature and intensity of
the ions present greatly varies as the collision energy is
changed from a very low value (where the molecular ion
is greatly predominant and little or no fragmentation is
observed) to increasingly higher ones (where the intensity of
the molecular ion gradually decreases and those of fragments
increase), no individual spectrum, recorded at a specific
value of collision energy, is fully representative of the
mass spectrometric behavior of the examined compound.
To overcome this limitation, the displayed fragment spectra
are ‘reconstructed’ from the entire set of fragment spectra,
to obtain a global representation of the fragmentation of
the selected precursor over the entire analysed range of
collision energy (typically from 0 to 2.5–4.0 eVCM). This is
simply accomplished by summing all individual normalized
spectra (thus compensating for the fluctuations in their
absolute intensities) into a single spectrum, which is in
turn normalized to the intensity of its most intense ion
to yield a final result which is referred to in this paper as
the ‘integrated’ spectrum. In this spectrum, all peaks with
intensities above 1% of the most intense ion or, if of lower
°
36 h at 40 C. All reaction mixtures were analysed by mass
spectrometry without prior separation or purification.
Mass spectrometric measurement
Mass spectrometric analyses were performed on an Applied
Biosystems API365 triple-quadrupole instrument (Applied
Biosystems, Monza, Italy), equipped with an ESI source
and operated in the positive-ion mode according to the
manufacturer’s instructions, essentially as reported.40–42
Briefly, 50-µl aliquots of the unprocessed reaction mixtures
were diluted 1 : 100 with the electrospray ionization solvent,
composed of a 1 : 1 v/v mixture of isopropyl alcohol and
water, containing 0.1% of formic acid to enhance ionization.
The orifice and ring potentials of the mass spectrometer
were optimized either to minimize in-source fragmentation
of some compounds (typical values were OR D 15 V and
RNG D 150 V or OR D 5 V and RNG D 50 V) or to purposely
generate molecular fragments for further ion analysis (typical
values were OR D 50 V and RNG D 350 V).
The source and fragment ion MS/MS spectra were
obtained in the multi-channel average mode over the
infusion time of the sample solution, until the signal/noise
ratio of the relevant signals was judged adequate. Ten
samples were acquired for each unit of m/z (step size of
Copyright  2006 John Wiley & Sons, Ltd.
J. Mass Spectrom. 2006; 41: 1578–1593
DOI: 10.1002/jms

ERMS of glutathione metaboloma metal conjugates
1581
intensity, with a clearly defined ‘peak shape’ (width <3 u;
width at half-height <0.5 u) were identified and were labeled
as candidate signals for interpretation.
and little or no fragmentation is observed, at increasingly
higher energies the intensity of the molecular ion gradually
decreases and those of the fragments increase, and in the
upper range the molecular ion is virtually absent and the
spectrum is dominated by intense, lower mass, further-
generation ion fragments, as exemplified in Fig. 1(a–c) for
protonated GSSG. Consequently, no individual spectrum,
recorded at a specific value of collision energy, is fully
representative of the richness and complexity of the mass
spectrometric behavior of the examined compound over
the range of explored collision energies. To overcome this
limitation, the fragment spectra in this paper (except for those
for which a specific value of collision energy is indicated)
are ‘reconstructed’ from the entire set of acquired fragment
spectra, to obtain a global representation of the fragmentation
of the selected precursor over the entire analysed range
of collision energy (typically from 0 to 2.5–4.0 eVCM).
This is simply accomplished by summing all individual
normalized spectra (thus compensating for fluctuations in
their absolute intensities) into a single spectrum, which
is in turn normalized to the intensity of its most intense
ion (‘Experimental’). The final results of this elaboration is
displayed as Fig. 1(d), which is referred to in this paper as
the ‘integrated’ fragment spectrum.
The fragment spectra of the bis-S-glutathionyl–metal (II)
conjugates of zinc and mercury are reported as examples
in Fig. 2(a) and (b). As apparent, the fragment spectra of
the GSH conjugates of several divalent heavy metals show
several similarities as well as a few prominent differences.
All yield as main fragments those due to loss of one neutral
GSH unit, formulated as the mono-S-glutathionyl–metal (II)
cation [GS–metalC], protonated glutathione [GSHŁHC] and
the fragments due to decomposition of the GSH portion
by loss of pyroglutamic acid [MH ꢀ 129]C and of glycine
[MH ꢀ 75]C. The array of fragments deriving from the
protonated bis-thiolato–metal (II) conjugates of the GSH
is very similar to that observed for GSSG (Fig. 1(d)), which
strengthens confidence that their chemical structure is that
of a peptide disulfide with the heavy metal atom interposed
between and chemically bound to the two sulfur atoms of
the disulfide bridge.
As for the differences between metals, zinc is the only
metal the conjugate of which shows a more intense [GSHŁHC]
fragment than the [GS-metalC] fragment, while only the
mercury (II) conjugate shows a [GSC] fragment (m/z 306)
with intensity very close to that of the [GSHŁHC] fragment.
The fragmentspectraofthe source-generated[GS-metalC]
ions of mercury and cadmium are reported as examples in
Fig. 3(b) and (c) and are compared to that of protonated
GSH (Fig. 3(a)). Some striking differences in the nature of
the generated fragment ions are apparent, which are possi-
bly underlying those in the chemical affinity of the metals
towards the ligand groups present in the GSH molecule:
the CySH thiol, the two carboxyl groups, the two peptide
bonds and the terminal amino function. Although a detailed
discussion of the differences among metals is beyond the
scope of this particular study, it is worth noting that none
of GSH–metal conjugates yields a ‘naked’ metalC ion in the
fragment spectra within the wide range of collision energies
The intensity of each ion signal present in the fragment
spectra was calculated on the basis of area, rather than height,
by employing the same front and back peak ‘boundaries’
(high and low m/z ends) at all values of collision energy,
irrespective of actual peak intensity or possible shape
deformation due to defocussing. The peak ‘boundaries’ for
each ion signal were determined from the peak shape of the
‘reconstructed’ spectrum (vide supra) as 5% of the maximum
intensity of the peak.
The proportion of integrated ion current assigned to
identified peaks was in all spectra >75%, most of the
discarded ion current being due to ‘trimming’ of the low
and (less frequently) high m/z tail of the most intense (and
wider) signals or to several fragments of very low intensity
which were not considered.
Elaboration of fragment ion appearance curves
To obtain the appearance curves of the fragment ions – which
describe as the relative intensity of each fragment ion as a
function of the energy imparted to the precursor ion by
collision with the target gas – the intensity of each identified
ion was calculated, at each value of collision energy, as a
percentage of the sum of all ions, including the surviving
precursor. The appearance curve of each ion was obtained
by plotting its relative intensity as a function of the center-
of-mass collision energy.
Calculation of onset energies
The obtained appearance curves of the fragments in the
MS/MS spectra of each examined compound were employed
to calculate the onset energies. This was accomplished by
interpolating the rising end of the curve with a straight
line and extrapolating the resulting equation to the point
where the intensity was zero. The obtained value, which
is expressed in electronvolt per molecule, is a rather crude
surrogate of the true threshold energy, i.e. of the minimum
energy required to generate a defined chemical species
(a fragmentation) from the selected precursor ion and to
observe it within the time frame allowed by the instrument.
For this reason, since the precursor ions of different analysed
molecules may have a different content of internal energy
upon electrospray ionization (which has not been measured
in the reported experiments), the onset energies for the first-
generation products are calculated relative to an arbitrarily
set value of zero of the mass-selected precursor ion, while
those of further-generation fragments (determined as such by
the order of onset energies and by chemical knowledge) are
calculated as the energy difference between the ‘onset energy’
of the considered fragment ion and that of its immediate
precursor.
RESULTS
The fragment spectra of all examined compounds were, as
expected, strongly dependant on the value of the collision
energy: at very low values the molecular ion is predominant
Copyright  2006 John Wiley & Sons, Ltd.
J. Mass Spectrom. 2006; 41: 1578–1593
DOI: 10.1002/jms

1582
F. M. Rubino et al.
Figure 1. Fragment mass spectra of protonated glutathione disulfide recorded at 5, 25 and 45 eV (laboratory frame) and integrated
spectrum obtained by summing 51 individual spectra acquired between 0 and 50 eVlab, corresponding to 0 and 2.2 eVcenterꢀofꢀmass
collision energy). The inset displays the disulfide fragments at m/z 306 (GSC) and 308 (GSHŁHC).
Copyright  2006 John Wiley & Sons, Ltd.
J. Mass Spectrom. 2006; 41: 1578–1593
DOI: 10.1002/jms

ERMS of glutathione metaboloma metal conjugates
1583
Figure 2. Integrated fragment mass spectra of (upper) protonated bis-S-glutathionyl–zinc (II) (spectra acquired between 0 and
1.4 eVcenterꢀofꢀmass collision energy, corresponding to 0 and 35 eVlab); (lower) protonated bis-S-glutathionyl–mercury (II) (spectra
acquired between 0 and 1.7 eVcenterꢀofꢀmass collision energy, corresponding to 0 and 52 eVlab).
experimented with. The most important difference among
the fragment spectra of the GS-metalC ions is that, of the four
examined heavy metal conjugates, only that of mercury (II)
yields a fragment ion at m/z 177 (25% of the total ion current),
which is assigned to oxidized cysteinylglycine (CSCyGlyOH)
and a weaker one at m/z 306 (2% of the total ion current), cor-
responding to oxidized GSH (ꢀGluCySCGlyOH). This means
that a mercury atom in its elemental state has been detached
from a RS–HgC species, thus generating a biothiol ion in an
oxidized state. Moreover, a very weak fragment at m/z 379
(0.2% of the total ion current) is observed, which is due to loss
of pyroglutamic acid directly from the m/z 508 precursor.
The resulting partial fragmentation pathway of protonated
bis-S-glutathionyl-mercury (II) is depicted in Scheme 2.
Occurrence of this fragmentation process has potential
toxicological significance, which will be detailed in the fol-
lowing discussion. An ERMS study was therefore performed
on the mono-glutathionyl–mercury (II) cation, as well as on
Scheme 2. Partial CAD fragmentation of bis-S-glutathionyl-mer-
cury (II).
Copyright  2006 John Wiley & Sons, Ltd.
J. Mass Spectrom. 2006; 41: 1578–1593
DOI: 10.1002/jms

1584
F. M. Rubino et al.
Figure 3. Integrated fragment mass spectra of: (upper) protonated glutathione (spectra acquired between 0 and 3.0 eVcenterꢀofꢀmass
collision energy, corresponding to 0 and 36 eVlab); (middle) mono-S-glutathionyl–cadmium (II) (spectra acquired between 0 and
2.8 eVcenterꢀofꢀmass collision energy, corresponding to 1 and 45 eVlab); (lower) mono-S-glutathionyl–mercury (II) (spectra acquired
between 0 and 2.7 eVcenterꢀofꢀmass collision energy, corresponding to 0 and 52 eVlab).
several thiols and disulfides of the GSH metaboloma contain-
ing ꢀ-linked glutamic acid residues, to better characterize the
observed processes through the measurement of the ‘onset
energy’ of the appearance curves of the relevant fragments.
The obtained results are exemplified by those of the three
key fragments of mono-glutathionyl–mercury (II), shown in
Fig. 4. As described in the ‘Experimental’ section, the ‘onset
energy’ extrapolated from the rising end of the curves is
related (although with strong limitations connected to the
nonideal energetic behavior of electrosprayed ions and to
the possible discrimination in the confinement and detec-
tion of fragment ions in the employed triple-quadrupole
instrument48–50) to the minimum energy necessary to gen-
erate a fragment from its precursor. It is apparent that the
fragments with a lower value of the ‘appearance energy’ are
formed (i.e. they are recorded in the spectra, irrespective of
their intensity) before those with a higher value, although
they can further decompose to other products, so that their
recorded intensity in the spectrum may be very low, as in
the case of m/z 379.
Copyright  2006 John Wiley & Sons, Ltd.
J. Mass Spectrom. 2006; 41: 1578–1593
DOI: 10.1002/jms

ERMS of glutathione metaboloma metal conjugates
1585
Copyright  2006 John Wiley & Sons, Ltd.
J. Mass Spectrom. 2006; 41: 1578–1593
DOI: 10.1002/jms

1586
F. M. Rubino et al.
Table 1. Loss of a ꢀ-linked residue of glutamic acid as the pyroglutamic acid neutral in several compounds
related to the glutathione metaboloma, each characterized by ERMS measurement of the onset energy (E0)
of the appearance curve of the relevant fragments
Precursor
(m/z)
E0
(eV)
Fragment
(m/z)
E0
(eV)
E0
(eV)
Compound
MHC
a
a
ꢀ-GluCySH
ꢁꢀ-GluCySꢂ2
251
499
251
499
370
308
322
613
484
427
815
686
508
306
122
370
241
179
193
484
355
298
686
557
379
177
0.41 š 0.06
0.43 š 0.02
0.77 š 0.02
0.35 š 0.03
0.35 š 0.05
0.34 š 0.03
0.67 š 0.04
0.35 š 0.02
0.34 š 0.02
0.53 š 0.02
0.81 š 0.04
0.58 š 0.03
0.41 š 0.06
0.43 š 0.02
0.34 š 0.03
0.35 š 0.03
0.35 š 0.05
0.34 š 0.03
0.34 š 0.05
0.35 š 0.02
0.34 š 0.02
0.19 š 0.03
0.81 š 0.04
0.22 š 0.06
0.43 š 0.02
a
GSH
GS-Me
GSSG
308
322
613
a
a
0.34 š 0.03
a
GS-SCy
GS-Hg-SG
427
815
a
0.34 š 0.02
GS-HgC
508
a
0.36 š 0.06
a For the calculation of the energy differences and uncertainties, the onset energy of the source-generated
precursor ion is arbitrarily considered as 0.00 š 0.00 eV.
The onset energies of the fragments in the spectra of
mono-S-glutathionyl-mercury (II) collected in Fig. 4 allow the
construction of the energy diagram depicted in Fig. 5, which
shows the partial decomposition processes undergone by
mono-S-glutathionyl-mercury (II) (CHgSG; m/z 508) finally
yielding m/z 177 through the two competing pathways of
Scheme 2, entailing stepwise loss of neutral pyroglutamic
acid and elemental mercury.
As is apparent, the energy required to detach a pyrog-
lutamic acid neutral from mono-S-glutathionyl-mercury (II)
(m/z 508) to generate m/z 379 (0.85 eV) is higher than that
required to generate m/z 177 from m/z 306 (0.22 eV) by loss of
neutral pyroglutamic acid. On the contrary, while 0.36 eV are
required to detach a neutral Hg0 from mono-S-glutathionyl-
mercury (II) (m/z 508) to yield m/z 306 (GSC; formally, a
half molecule of ‘oxidized’ GSH), any m/z 379 formed spon-
taneously undergoes oxidative loss of Hg0 to yield m/z 177,
since the potential of the m/z 379 fragment (0.85 eV) is 0.27 eV
higher than that of the m/z 177 product (CSCyGly; formally,
a half molecule of oxidized cysteinylglycine).
several compounds related to the GSH metaboloma. The
obtained values are collected in Table 1.
When the decomposition occurs only once from the
molecular precursor (in the cases of protonated GSH,
ꢀ-glutamyl-cysteine and GS-Me), the value is calculated as
the difference between the onset energy of the [MH ꢀ Pyr]C
fragment and that of the MHC ion, which is arbitrarily
considered as zero. This procedure introduces a systematic
bias in the calculation of the absolute energy values, since it
does not take into account the effect of focussing aberrations,
space charge limitations and contact potentials, which
influence the true energy content of the precursor ion. When
two such fragmentation processes occur consecutively, as in
ꢀ-glutamyl-cysteine disulfide, GSSG and bis-glutathionyl-
mercury (II), the energy is calculated as the difference
between the appearance energy values of the two fragments
[MH ꢀ Pyr]C and [MH ꢀ 2Pyr]C. The obtained values are
around 0.35 eV, within a few tens of meV to one another, in
compounds carrying either one or two ꢀ-connected residues
of glutamic acid.
According to studies in the literature, the loss of the
ꢀ-glutamyl group as pyroglutamic acid and formation of
the [MH ꢀ Pyr]C fragment may proceed as depicted in
Scheme 3.51–53 A nucleophilic attack of the free amino group
of glutamic acid occurs on the ꢀ-carbonyl involved in the
(pseudo) peptide bond to the cyst(e)ine residue, thereby
releasing a cyclized unit of pyroglutamic acid (Pyr). The
two fragments initially share a proton, which is captured
by the much more basic dipeptide fragment, to yield the
protonated [MH ꢀ Pyr]C species, while the complementary
protonated species (m/z 130) is observed only at higher
values of collision energy, above approximately 0.86 eVCM
(data not shown). When two residues of glutamic acid are
present, stepwise loss of the two residues occurs, as observed
in GSSG and in bis-S-glutathionyl-mercury (II).
Although the absolute values of the energy involved in
loss of the ꢀ-linked glutamic acid residue from the different
examined precursors are not accurate, what is immediately
noticed is that the energies involved in the corresponding
process in the fragmentation of mono-glutathionyl-mercury
(II)C, namely, m/z 508 ! 379 (0.86 eV) and m/z 306 ! 177
(0.27 eV) are much higher in one case and slightly lower in the
other than that of 0.35 eV measured for several protonated
GSH compounds.
In particular, the higher energy required to remove
the N-terminal ꢀ-glutamic residue of GSH from the mono-
glutathionyl-mercury (II)C ion may be explained as the
consequence of a strong ionic interaction of the deprotonated
carboxylic group of the N-terminal ꢀ-glutamic residue to the
cationic mercury (II)-thiolate group, which in turn hampers
the conformation required for the fragmentation to occur.
Therefore, mono-glutathionyl-mercury (II)C ions with the
The onset energy for occurrence of this reaction was
measured from the appearance curve of the fragments in
Copyright  2006 John Wiley & Sons, Ltd.
J. Mass Spectrom. 2006; 41: 1578–1593
DOI: 10.1002/jms

ERMS of glutathione metaboloma metal conjugates
1587
Scheme 3. Postulated decomposition mechanism for loss of the C-terminal, ꢀ-linked glutamic acid residue from the peptides of the
glutathione metaboloma.
Scheme 4. Postulated decomposition mechanism for loss of the C-terminal, ꢀ-linked glutamic acid residue from
mono-S-glutathionyl-mercury(II).
cationic mercury-thiolate bound to a carboxylate group of the
peptide in the zwitterionic form (either the glycine carboxyl
or that of the ˛-glutamyl amino-carboxylic residue, which is
possibly one of the lowest-energy conformations of the m/z
508 ions) must convert to the more energetic conformation
with the ˛-glutamyl amino-carboxylic group unbound and
then from the zwitterionic to the neutral form, prior to
decomposing to m/z 379 through loss of the ꢀ-glutamyl
group. Even in the absence of confirmation by quantum
mechanical calculations (which are extremely complicated on
account of the large number of atoms in the system, its great
conformational flexibility and the unavoidable presence of
the highly polyelectronic atom of mercury), Scheme 4 shows
the individual steps of the isomerization of m/z 508 necessary
to yield the product at m/z 379 according to the tentatively
postulated mechanism.
the reaction, in analogy to the decomposition process
depicted in Scheme 3. That this may really be the case can
be confirmed by considering that, in the fragment spectrum
of protonated GSSG, both the m/z 308 (GSHŁHC) and the
m/z 306 (GSC) first-generation fragments further decompose
by loss ofŁneutral pyroglutamic acid to yield the m/z 179
(HSCyGly HC) and m/z 177 (CSCyGly) ions, respectively. As
deduced from the ‘integrated’ spectrum of Fig. 1(d), the m/z
177/306 intensity ratio (6.1%/0.2% D 30.5) is much higher
than the 179/308 one (1.3%/0.6% D 2.2): this means that loss
of neutral pyroglutamic acid is much easier from m/z 306
than from m/z 308.
Moreover, when considering the integrated fragment
spectrum of mono-glutathionyl-mercury (II) of Fig. 3, the
intensity ratios of the two processes involving loss of neutral
mercury from RS-HgC precursors to yield RSC species are
even more different: the ratio of m/z 306/508 is 0.08, while
that of m/z 177/379 is 123. This very large difference, which
is consistent with the energy diagram of Fig. 5, thus suggests
that the ꢀ-glutamic acid residue of GSH may have a crucial
On the contrary, the slightly lower energy required
to generate m/z 177 from m/z 306 can be explained by
considering that the formal sulfenium group may act as an
electrophile to the ꢀ-glutamyl carbonyl, thereby facilitating
Copyright  2006 John Wiley & Sons, Ltd.
J. Mass Spectrom. 2006; 41: 1578–1593
DOI: 10.1002/jms

1588
F. M. Rubino et al.
fragments are, with the exception of disulfide amino acids
in which they are prominent, generally of weak intensity,
especially when the molecule has lower-energy processes
available (as in the case of GSSG, in which the loss of the
ꢀ-linked glutamic acid residue is that of the lowest energy
and most relevant decomposition, as seen before). Although
the detailed mechanism of their formation in the gas phase
is not demonstrated in sufficient detail, it is nevertheless
conceivable that the relative intensity of the two species
may be related to the electrochemical redox potential of the
corresponding thiol–disulfide pair. The ERMS study also
showed that the intensity ratio of the RSC and RSHŁHC
species is not constant over the range of collision energies,
and may be increasing (as exemplified in the graph of
Fig. 6, which shows the appearance curves of the disulfide
fragments of homodimeric HSCyGly disulfide), decreasing
or even bell-shaped (not shown). Therefore, to obtain a
meaningful value of the [RSC]/[RSHŁHC] ion ratio, it needs to
be calculated not with reference to an individual, arbitrarily
selected value of collision energy, but considering all the
relevant range of collision energy, as shown in Table 2.
A comparison of the [RSC]/[RSHŁHC] ion ratio – which
closely represents in the fragment spectra of the disulfides
the equilibria of the thiol/disulfide redox pair – to the
electrochemical potentials measured in aqueous solution is,
however, possible only for a few compounds, the reduction
electrochemical potentials of which have been recently
measured by other physico-chemical techniques54,55 and are
reported in Table 2 (column 8). The plot of Fig. 7 shows that
a fair agreement exists between the solution electrochemical
potentials of three out of four compounds reported in
the literature54 and the ERMS-measured [RSC]/[RSHŁHC]
intensity ratios, the noticeable exception being homo-cystine.
Although the linear relationship is obtained on only three
out of four available values (the outlying compound being
discarded from further evaluation on the basis of a structural
difference to the others, the influence of which on disulfide
bond fission had previously been briefly rationalized40), the
quality of the calculated regression allows the calculation
of the electrochemical reduction potentials of the remaining
thiols by linear extrapolation, as listed in the last column of
Table 2.
Figure 5. Energy diagram for the fragmentation processes
leading to the ions at m/z 379 (CHgSCyGly), m/z 306 (GSC)
and m/z 177 (CSCyGly) from mono-glutathionyl-mercury (II)
(CHgSG; m/z 508).
role in stabilizing the bond of the mercury (II) ion in the
mono-coordinated conjugate, hampering loss of the metal as
a neutral. In other words, loss of neutral mercury from the
RS-HgC species in the gas phase may be considered as a
formal redox reaction implicitly taking place when a neutral
mercury atom is released by a formal two-electron reduction
to the elemental state and the concurrent oxidation of the
thiolate to the sulfenium form in a gas-phase collisionally
activated dissociation process. The obtained results therefore
suggest that mercury (II) is able to more efficiently oxidize
the thiolate of HSCyGly than that of GSH to the respective
sulfenium species or, in a complementary way, that HSCyGly
is a stronger reducing agent of mercury (II) than is GSH.
This point can be again demonstrated through an
ERMS study of several homodimeric thiol disulfides of
the GSH metaboloma. CAD of their protonated molecules
shows formation of both [RSC] and [RSHŁHC] species,
which are in the context of this paper referred to as the
‘disulfide fragments’. As previously observed,40 the disulfide
Table 2. Disulfide ions in the fragment spectra of protonated homodimeric thiol disulfides of the glutathione metaboloma and
corresponding solution electrochemical redox potentials
IRS
(%)
C
IRSH H
[RSC]/
[RSHŁHC]
Ł
C
MHC
RSC
RSHŁHC
(%)
E0
E0
a
b
Compound
CyS-Scy
241
297
269
325
381
355
499
613
120
148
134
162
190
177
249
306
21.1
8.9
122
150
136
164
192
179
251
308
9.6
1.8
24.7
5.0
0.2
1.6
0.6
0.6
2.20
4.94
0.66
6.00
50.00
9.75
5.33
0.33
ꢀ0.247
ꢀ0.267
ꢀ0.196
ꢀ0.248 š 0.003
ꢀ0.266 š 0.003
ꢀ0.220 š 0.003
ꢀ0.271 š 0.003
ꢀ0.319 š 0.003
ꢀ0.282 š 0.003
ꢀ0.268 š 0.003
ꢀ0.205 š 0.003
PenS-SPen
hCyS-ShCy
NACyS-SCyNA
NAPenS-SPenNA
ꢁSCyGlyꢂ2
16.3
30.0
10.0
15.6
3.2
ꢁgGluCySꢂ2
GSSG
0.2
ꢀ0.205
a Solution electrochemical reduction potentials are reported from Rabenstein, 1993.54,55
b Solution electrochemical potentials recalculated from the equation displayed in Fig. 7.
Copyright  2006 John Wiley & Sons, Ltd.
J. Mass Spectrom. 2006; 41: 1578–1593
DOI: 10.1002/jms

ERMS of glutathione metaboloma metal conjugates
1589
Figure 6. (left) Appearance curves of the disulfide fragments of protonated cysteinylglycine disulfide. (right) Intensity ratio of the m/z
177/179 disulfide fragments as a function of the center-of-mass collision energy.
have in explaining the mechanism of the cellular toxicity of
mercury will be discussed below.
DISCUSSION AND CONCLUSIONS
The use of triple-quadrupole tandem mass spectrometry
to record the fragment spectra of electrosprayed proto-
nated GSH metaboloma thiols, disulfides and conjugates
at several increasing values of collision energy has allowed
highlighting the occurrence of several important fragmen-
tation processes of the protonated precursors, expecially
those occurring consecutively from the precursor ion at rela-
tively high values of collision energy. However, the current
unavoidable experimental limitations in measuring onset
potentials (mainly due to a poor control of the internal energy
of the precursor ion and of the detection of fragment ions48–50
)
Figure 7. Plot of log([RSC]/[RSHŁHC]) versus solution
electrochemical reduction potentials of thiol disulfides. The ion
ratios are calculated from the intensities of Table 2; the
solution electrochemical reduction potentials are reported from
Rabenstein, 1993.54 The linear regression is calculated from
the data of the disulfides for which solution electrochemical
potentials have been reported: those of CySH, PenSH and
GSH (boxed in the plot). The confidence limits of the
regression are shown as the narrow lines. The solution
electrochemical potentials for the other disulfides have been
recalculated from the displayed equation. The data for the
outlier homo-cystine is reported as the closed square.
make the obtained results rather crude surrogates of reac-
tion energies and discourage a detailed comparison of the
measurements performed – and especially of the observed
differences – among the different examined compounds.
The fragmentation pathways of the analysed GSH
metaboloma closely mimic well-known biochemical transfor-
mations of GSH metabolites into the excreted end products,
as in particular, exemplified by the energy-resolved mass
spectrometric study of the process leading to loss of the C-
terminal, ꢀ-linked residue of glutamic acid as pyroglutamic
acid (or 5-oxo-proline). In analogy with the mass spectro-
metric fragmentation of GSSG, the fragmentation of the
GSH-mercury conjugate closely resembles that occurring by
enzymatic cleavage in ex vivo preparations of rabbit renal
tubules,56 mainly represented by removal of the N-terminal
ꢀ-glutamic acid residue by ꢀ-glutamyl transferase. It is worth
noting that 5-oxo-proline is a known catabolite of the GSH
metaboloma, which is excreted in the urine in large amounts
in human subjects with a congenital defect of the oxoproli-
nase enzyme.57,58
In particular, according our measurements, HSCyGly has
the second largest [RSC]/[RSHŁHC] intensity ratio among
the examined homodimeric disulfides and the second lowest
(calculated) reduction potential, by 0.078 V lower (i.e. more
reducing) than GSH. The results of this measurement may
explain the observed differences in the energy required
for ‘reductive’ detachment of elemental mercury from the
fragment ions containing GSH (m/z 508) or HSCyGly (m/z
379) as thiol ligands. The consequences that these results may
That the mercury(II) conjugates of the GSH metaboloma
are substrates of the same enzymes that proteolytically
Copyright  2006 John Wiley & Sons, Ltd.
J. Mass Spectrom. 2006; 41: 1578–1593
DOI: 10.1002/jms

1590
F. M. Rubino et al.
process GSH and GSSG is very likely on the basis of
several lines of evidence, which all stress the great similarity
(molecular homology or ‘mimicry’) of their molecular shape
and charge distribution of metal (II)–thiolate conjugates
with those of the corresponding metabolites of organic
xenobiotics and of physiologically occurring analogues.59–61
It has recently been stressed that the application of this
concept to the understanding of the biochemical behavior of
xenobiotics may be considerably flawed if the geometry and
charge distribution of the individual molecules is not taken
into deep consideration.62
ERMS data yield additional confidence to biochemical
observations and measurements, valuable since they are
coming from an entirely different perspective. To measure
appreciably close onset energies for the loss of pyroglutamic
acid from different precursors, including the ones containing
a heavy metal substituent, conceivably means that the con-
formational requirements for this reaction are substantially
independent on the presence of substituents located suffi-
ciently far from the reacting centre. Transposing the mass
spectrometric results (i.e. a reactivity study of molecules
performed on isolated gas-phase ions) to their aimed bio-
chemical meaning of reactivity in water solution, this finding
stands with the indirect evidence56 that biochemically formed
or exogenously administered bis-S-glutathionyl-mercury (II)
is also a substrate of ꢀ-glutamyl-transferase, able to undergo
enzymatic loss of the pyroglutamic acid residue in full anal-
ogy with GSH adducts of xenobiotics, here represented by
GS-Me, which are catabolized in vitro and ex vivo to the
finalS-methylcysteine and S-methylmercapturic acid urinary
metabolites.
Another important issue in the biochemical reactions
of the GSH metaboloma and its interaction with thiol-
seeking heavy metals, which could be tackled with the
use of energy-resolved tandem mass spectrometric exper-
iments, is represented by the thiol–disulfide redox reactions.
Heterolytic fission of the disulfide bond yields the comple-
mentary sulfenium and protonated thiol fragments in a series
of homodimeric disulfides of the GSH metaboloma. The
[RSC]/[RSHŁHC] intensity ratio was found to be inversely
proportional to the value of the electrochemical reduction
potential of each of the thiols for which such values had been
measured and reported in the literature. This trend, although
limited to only three compounds, is logically compatible
with the underlying expected chemicalCbehavior, since more
reducing thiols have also a higher [RS ]/[RSHŁHC] ratio,55
i.e. their unimolecular decomposition favors the formation
of the ‘oxidized’ versus the ‘reduced’ form.
(V) to Sb (III) in anti-parasitary drugs than CySH, GSH and
tripanothione,63 the efficacy of the three GSH metaboloma
thiols being in the same order of their electrochemical
potentials in solutions and of the ratios of the disulfide
fragments measured by mass spectrometry. Also, a GSH
oxidase activity was long sought for the enzyme ꢀ-glutamyl-
transpeptidase, which generates HSCyGly from GSH,64 and
was finally discarded as an artefact due to rapid oxidation of
the released dipeptide.65
All these mass spectrometric results put together allow
highlighting the facets of the still poorly explored chemical
behavior of the metal (II) conjugates of the GSH metaboloma
peptides and hint at unforeseen mechanisms for their
interaction with the biological environment, ultimately
leading to cellular and organ toxicity. Oxidative stress
is a general cellular condition wherein overproduction of
oxidizing species leads to a depletion of intracellular redox-
buffering systems to a point that the cellular environment
cannot cope and therefore triggers apoptotic signals or the
cell undergoes necrosis.66 The general mechanism for heavy-
metal-induced cellular toxicity has commonly been assigned
to their interaction with essential protein thiol groups,
which may impair their catalytic activity, as demonstrated
at the atomic scale for the inhibition of ALA-synthase by
lead.67 Nevertheless, the diversity of clinical presentation
of the toxicological effects of heavy metals15–20 points at
the existence of metal-specific mechanisms, which may be
justified on the basis of the differences in their chemical
behavior.68–70
Among the four heavy metals (zinc, cadmium, mercury
and lead), the GSH conjugates of which have been studied
by mass spectrometry, mercury is the one with the highest
ionization energy and the only one to have a positive redox
potential, which fact may explain its ability to be detached
from metal (II)-thiolates as a neutral atom through formal
oxidation of the sulfur atom to a sulfenium species, as
has been observed in its fragment mass spectra. In the gas
phase this process may possibly be hampered by formation
of a strong salt bridge with the carboxylate functions of
GSH; therefore loss of the ꢀ-glutamyl residue, which is
possibly the strongest binding group, removes the inhibition
and enhances loss of the metal as a neutral, thus possibly
justifying also its lower electrochemical potential.
Toxicological experiments with synthetic Hg-biothiol
conjugates performed on tubular kidney preparations
showed that the toxicity of exogenously administered
Hg(SCyGly)2 towards renal proximal tubule cells is higher
than that of either ꢁGSꢂ2Hg or of ꢁCySꢂ2Hg. Inhibition of the
degradation of ꢁGSꢂ2Hg to Hg(SCyGly)2 by treatment with
a specific inhibitor of ꢀ-glutamyl-transpeptidase (GGT), the
catabolic enzyme of the GSH-mercapturate pathway, did
not decrease Hg uptake, but decreased its cytotoxicity.56
Also another experiment, performed in cell culture, showed
a slightly higher toxicity of mercury in the presence of
HSCyGly than of GSH.71
The electrochemical reduction potentials of some disul-
fide/thiol redox pairs for which literature data is not avail-
able could be calculated on the basis of the measured ion
ratios and, in particular, HSCyGly resulted as the strongest
reducing thiol of the natural GSH metaboloma.
That the employed approach may have a value in
determining the order of the reduction potentials of thiol
disulfides is also suggested by interpretation of some
scattered chemical and biochemical evidence. As far as
HSCyGly is concerned, it was recently found that this
thiol dipeptide is a more effective reducing agent of Sb
All this evidence, considered together, suggests that
mercury (II), but none of the other considered heavy
metals, is able to trigger catalytic depletion of the cellular
GSH pool by a catalytic mechanism acting on the metal
Copyright  2006 John Wiley & Sons, Ltd.
J. Mass Spectrom. 2006; 41: 1578–1593
DOI: 10.1002/jms

ERMS of glutathione metaboloma metal conjugates
1591
measurement of their chemical equilibria with ligands, such
as enzyme substrates, receptor ligands and inhibitors,80 as
well as of much larger bio-structures, such as antibodies,
regulatory proteins and organelles.81 Our approach, based
on the use of ERMS data to approach problems in mech-
anistic toxicology, aims to apply the currently available
power of tandem mass spectrometry beyond its now-long-
recognized value in the structural determination, identifi-
cation and quantitation of biologically and toxicologically
relevant metabolites of endogenous, xenobiotic and envi-
ronmental chemicals. The intrinsic limitations of the use of
analytically oriented instrumentation for physico-chemical
measurements, such as threshold measurements, have been
thoroughly discussed and the advantages for using spe-
cial dedicated equipment, such as guided ion beam mass
spectrometers, for such measurements have been stressed.
Nevertheless, a few examples of the wealth of information
that ERMS can yeld not only to interpret fragmentation mech-
anisms, but also to gain insight into the chemical behavior of
biomolecules have been forwarded.43
With all the limitations in mind on the far-from-optimum
quality of the obtained onset energies, which only grossly
approximate the true thermochemical quantities needed to
support any deeper insight, our results allow understanding,
at least on a semiquantitative basis, the different mechanisms
of metal toxicity and stress the value of mass spectrometry
not only as an analytical technique for environmental and
biological monitoring of metal pollution, but mainly also as
a powerful tool to investigate the toxic mechanism of the
different metals from the point of view of their transport
forms within biological systems and of induced perturbation
within the cellular environment.82–84
Scheme 5. Catalytic cycle for the depletion of the cellular
glutathione pool by mercury, proposed on the basis of mass
spectrometric behavior and biochemical evidence.
(II)-thiol conjugates and involving as a necessary step the
enzymatic removal of the glutamic acid residue to yield
a mercury (II)-cysteinyl-glycine conjugate, as depicted in
Scheme 5. Enzymatic removal of the ꢀ-glutamyl residue of
bis-glutathionyl-mercury (II) by the enzyme GGT generates
cysteinylglycyl–mercury (II) conjugates, which are able to
release the metal in its reduced form within the cellular
environment, themselves transforming to the oxidized
disulfides. This phenomenon triggers a catalytic mechanism
of oxidative stress,72–76 since the released elemental mercury
can thus be re-oxidized in the intracellular compartment by
several enzymes, such as catalase77 and re-enter the catalytic
GSH-depleting cycle. On the contrary, in the GSH conjugate,
the presence of the ˛-glutamyl carboxyl group stabilizes the
binding of mercury (II) to the thiolate ligand by a dipolar
interaction, thereby preventing its loss as a neutral metal
atom.
The proposed mechanism of oxidative stress through
depletion of the reduced GSH pool by catalytic mercury-
mediated oxidation is of general application to all cell
types, where biosynthesis, transport, intracellular compart-
mentation and interconversion of GSH metaboloma and
of xenobiotic conjugates occur with the same mechanism.
Any depletion of the reduced GSH pool of cells drifts their
cellular redox potential towards a more oxidizing status
(positive redox potential) and therefore trigger apoptosis or
necrosis.66,78 The other heavy metals, on the contrary, are able
to decrease only the cellular pool of reduced GSH by stoi-
chiometric formation of electrochemically inert, catalytically
inactive67,79 mercaptides.
Acknowledgements
The authors are very grateful to Prof. Peter B Armentrout (University
of Utah, USA) for helpful discussion and generous advice in the
preparation of this paper.
REFERENCES
1. Dickinson DA, Forman HY. Cellular glutathione and thiols
metabolism. Biochem. Pharmacol. 2002; 64: 1019.
2. Filomeni G, Rotilio G, Ciriolo MR. Cell signalling and the
glutathione redox system. Biochem. Pharmacol. 2002; 64: 1057.
3. Landino LM, Moynihan KL, Todd JV, Kennett KL. Modulation
of the redox state of tubulin by the glutathione/glutaredoxin
reductase system. Biochem. Biophys. Res. Commun. 2004; 314: 555.
4. Paolicchi A, Dominici S, Pieri L, Maellaro E, Pompella A.
Glutathione catabolism as signaling mechanism. Biochem.
Pharmacol. 2002; 64: 1027.
5. Kleinman WA, Komninou D, Leutzinger Y, Colosimo S, Cox J,
Lang CA, Richie JP Jr. Protein glutathionylation in human blood.
Biochem. Pharmacol. 2003; 65: 741.
6. Kuo-Ping H, Freesia LH. Glutathionylation of proteins by
glutathione disulfide S-oxide. Biochem. Pharmacol. 2002; 64: 1049.
7. Pastore A, Mozzi AF, Tozzi G, Gaeta LM, Federici G, Bertini E,
Lo Russo A, Mannucci L, Piemonte F. Determination of
glutathionyl-hemoglobin by cation-exchange high-performance
liquid chromatography. Anal. Biochem. 2003; 312: 85.
8. Casagrande S, Bonetto V, Fratelli M, Gianazza E, Eberini I,
Massignan T, Salmona M, Chang G, Holmgren A, Ghezzi P.
Glutathionylation of human thioredoxin: a possible crosstalk
between the glutathione and thioredoxin systems. Proc. Natl.
Acad. Sci. USA 2002; 99: 9745.
In conclusion, the use of biologically compatible ioniza-
tion techniques, such as electrospray, is now directed to
obtain gas-phase intact ions of supramolecular assemblies of
biological macromolecules, as such amenable to the direct
Copyright  2006 John Wiley & Sons, Ltd.
J. Mass Spectrom. 2006; 41: 1578–1593
DOI: 10.1002/jms

1592
F. M. Rubino et al.
9. Sengupta S, Wehbe C, Majors AK, Ketterer ME, DiBello PM,
Jacobsen DW. Relative roles of albumin and ceruloplasmin
in the formation of homocystine, homocysteine-cysteine-mixed
disulfide and cystine in circulation. J. Biol. Chem. 2001; 276: 46896.
10. Harada D, Naito S, Hiraoka I, Otagiri M. In vivo kinetic analysis
of covalent binding between N-acetyl-cysteine and plasma
protein through the formation of mixed disulfides in rats. Pharm.
Res. 2002; 19: 615.
11. Harada D, Naito S, Otagiri M. Kinetic analysis of covalent
binding between N-acetyl-cysteine and albumine through the
formation of mixed disulfides in human and rat serum in vitro.
Pharm. Res. 2002; 19: 1648.
12. Georgius Agricola. De Re Metallica [TRANSLATED FROM THE
FIRST LATIN EDITION OF 1556 With Biographical Introduction,
Annotations And Appendices Upon The Development Of Mining
Methods, Metallurgical Processes, Geology, Mineralogy & Mining
Law From The Earliest Times To The 16th Century. By Herbert Clark
Hoover and Lou Henry Hoover]. Dover Publications, 1950.
13. Ramazzini B. De Morbis Artificum [On The Workers’ Illnesses] 1713
edn, Engl Transl. by WilmerCave Wright, University of Chicago
Press: Chicago, 1940; 549.
14. Felton JS. The heritage of Bernardino Ramazzini. Occup. Med.
(Lond.) 1997; 47: 167.
15. Olanow CW. Manganese-induced parkinsonism and Parkin-
son’s disease. Ann. N.Y. Acad. Sci. 2004; 1012: 209.
16. Herrero Hernandez E, Discalzi G, Dassi P, Jarre L, Pira E.
Manganese intoxication: the cause of an inexplicable epileptic
syndrome in a 3 year old child. Neurotoxicology 2003; 24: 633.
17. Mascagni P, Consonni D, Bregante G, Chiappino G, Toffoletto F.
Olfactory function in workers exposed to moderate airborne
cadmium levels. Neurotoxicology 2003; 24: 717.
30. Urano T, Naganuma A, Imura N. Methylmercury-cysteinylgly-
cine constitutes the main form of methylmercury in rat bile. Res.
Commun. Chem. Pathol. Pharmacol. 1988; 60: 197.
31. Rabenstein DL, Isab AA, Reid S. A proton nuclear magnetic
resonance study of the binding of methylmercury in human
erythrocytes. Biochim. Biophys. Acta 1982; 696: 53.
32. Rabenstein DL, Isab AA. A proton nuclear magnetic resonance
study of the interaction of mercury with intact human
erythrocytes. Biochim. Biophys. Acta 1982; 721: 374.
33. Li Z-S, Lu Y-P, Zhen R-G, Szczypka M, Thiele DJ, Rea PA.
A
new pathway for vacuolar cadmium sequestration
in saccharomyces cerevisiae: YCF1-catalyzed transport of
bis(glutathionato)cadmium. Proc. Natl. Acad. Sci. USA 1997; 94:
42.
34. DeWitt JG. Analysis of phytochelatin-cadmium complexes
from
plant
tissue
culture
using
nano-electrospray
ionization tandem mass spectrometry and capillary
liquid chromatography/electrospray ionization tandem mass
spectrometry. J. Mass Spectrom. 1999; 34: 930.
35. El-Zohri MHA, Cabala R, Frank H. Quantification of
phytochelatins in plants by reverse-phase HPLC-ESI-MS-MS.
Anal. Bioanal. Chem. 2005; 382: 1871.
36. Navaza AP, Montes-Bayon M, LeDuc LL, Terry N, Sanz-
Medel A. Study of phytochelatins and other related thiols as
complexing biomolecules of As and Cd in wild type and
genetically modified brassica juncea plants. J. Mass Spectrom.
2006; 41: 323.
37. O’Hair RAJ, Styles ML, Reid GE. Role of the sulfhydryl group
on the gas-phase fragmentation reactions of protonated cysteine
and cysteine containing peptides. J. Am. Soc. Mass Spectrom. 1998;
9: 1275.
18. Vigliani EC, Baldi G. [Unusual epidemics of mercury poisoning in a
felt hat factory. 1949] (reprinted in: Med. Lav. 2001; 92: 391).
19. Wedeen RP. Were the hatters of New Jersey ‘‘mad’’? Am. J. Ind.
Med. 1989; 16: 225.
38. Reid GE, Simpson RJ, O’Hair RAJ. Leaving group and gas-
phase neighbouring group effect in the side chain losses from
protonated serine and its derivatives. J. Am. Soc. Mass Spectrom.
2000; 11: 1047.
20. Giavarini C. Gli anni del piombo [the years of lead]. Chim Ind
(Milano) 1990; 72: 1027.
39. Budi B, Sengupta S, DiBello PM, Kinter MT, Jacobsen DW. A
facile synthesis of Homocysteine-Cysteine mixed disulfide. Anal.
Biochem. 2001; 291: 303.
21. Tang W, Stearns RA, Bandiera SM, Yong Z, Raab C, Braun MP,
Dean DC, Pang J, Leung KH, Doss GA, Strass JR, Kwei GY,
Rushmore TH, Chiu S-HL, Baillie TA. Studies on cytochrome
p-450-mediated bioactivation of diclofenac in rats and in
human hepatocytes: identification of glutathione conjugated
metabolites. Drug Metab. Disp. 1999; 27: 365.
22. Scholz C, Dekant W, Voelkel W, Paehler A. Rapid detection and
identification of N-acetyl-L-cysteine thioethers using constant
neutral loss and theoretical multiple reaction monitoring
combined with enhanced product-ion scans on a lineari on
trap mass spectrometer. J. Am. Soc. Mass Spectrom. 2005; 16: 1976.
23. Gucˇek M, Makuc S, Mlakar A, Bericˇnik-Vrbovsˇek J, Marsel J.
Determination of glutathione in spruce needles by liquid
chromatography/tandem mass spectrometry. Rapid Commun.
Mass Spectrom. 2002; 16: 1186.
24. Guan X, Hoffman B, Dvivedi C, Matthees DP. A simultaneous
liquid chromatography/massspectrometry assay ofglutathione,
cysteine, homocysteine and their disulfides in biological
samples. J. Pharm. Biomed. Anal. 2003; 31: 251.
25. Bouligand J, Deroussent A, Paci A, Morizet J, Vassal G. Liquid
chromatography-tandem mass spectrometry assay of reduced
and oxidized glutathione and main precursors in mice liver.
J. Chromatogr. B 2006; 832: 67.
40. Rubino FM, Verduci C, Giampiccolo R, Pulvirenti S, Bram-
billa G, Colombi A. Characterization of the disulfides of bio-
thiols by Electrospray ionization and triple quadrupole tandem
mass spectrometry. J. Mass Spectrom. 2004; 17: 927.
41. Rubino FM, Verduci C, Giampiccolo R, Pulvirenti S, Bram-
billa G, Colombi A. Molecular characterization of homo-and
Heterodimeric mercury(II)-bis-Thiolates of some biologically rel-
evant Thiols by Electrospray ionization and triple quadrupole
tandem mass spectrometry. J. Am. Soc. Mass Spectrom. 2004; 15:
288.
42. Rubino FM, Pitton M, Brambilla G, Colombi A. Electrospray
ionization and triple quadrupole tandem mass spectrometry
study of some biologically relevant homo- and Heterodimeric
cadmium Thiolate Conjugates. J. Am. Soc. Mass Spectrom. 2006;
17: 1442.
43. Ho Y, Kebarle P. Studies of the dissociation of deprotonated
mononucleotides by energy resolved collision-induced
dissociation. Int. J. Mass Spectrom. Ion Processes 1997; 165/166:
433.
44. Klassen JS, Kebarle P. Collision-indiced dissociation threshold
energies of protonated glycine, glycinamide and some related
small peptides and peptide amino amides. J. Am. Chem. Soc.
1997; 119: 6552: 6563.
26. Ohsawa M, Magos L. The chemical form of the methylmercury
complex in the bile of the rat. Biochem. Pharmacol. 1974; 23: 1903.
27. Cherian MG, Vostal JJ. Biliary excretion of cadmium in rat. I.
Dose-dependent biliary excretion and the form of cadmium in
the bile. J. Toxicol. Environ. Health 1977; 2: 945.
45. Meister A, Anderson ME. Glutathione. Annu. Rev. Biochem. 1983;
52: 711.
46. Ham BM, Cole RB. Determination of bond dissociation energies
using Electrospray tandem mass spectrometry and a derived
effective reaction path length approach. Anal. Chem. 2005; 77:
4148.
28. Ballatori N, Clarckson TW. Inorganic mercury secretion into bile
as a low MW complex. Biochem. Pharmacol. 1984; 33: 1087.
29. Katoh M, Mizutani N, Keino H, Kashiwamata S.
A
low-
47. Busch KL, Glish S, McLuckey SA. Mass Spectrometry/Mass
Spectrometry: Techniques and Applications of Tandem Mass
Spectrometry. VCH Publication: New York, 1988.
molecular-weight cadmium-binding substance in human and
rat livers and human blood. Toxicol. Appl. Pharmacol. 1984; 73: 97.
Copyright  2006 John Wiley & Sons, Ltd.
J. Mass Spectrom. 2006; 41: 1578–1593
DOI: 10.1002/jms

ERMS of glutathione metaboloma metal conjugates
1593
48. Rodgers MT, Armentrout PB. Noncovalent metal-ligand bond
energies as studied by threshold collision-induced dissociation.
Mass Spectrom. Rev. 2000; 19: 215.
49. Armentrout PB. Kinetic energy dependence of ion-molecule
reactions: guided ion beams and threshold measurements. Int. J.
Mass Spectrom. 2000; 200: 219.
50. Armentrout PB. Mass spectrometry–notjustastructural tool: the
use of guided ion beam tandem mass spectrometry to determine
thermochemistry. J. Am. Soc. Mass Spectrom. 2002; 13: 419.
51. Dookeran NN, Yalcin T, Harrison AG. Fragmentation reactions
of protonated peptides. J. Mass Spectrom. 1996; 31: 500.
52. Harrison AG. Ion chemistry of protonated glutamic acid
derivatives. Int. J. Mass Spectrom. 2001; 210/211: 361.
53. Harrison AG. Fragmentation reactions of protonated peptides
containing glutamine or glutamic acid. J. Mass Spectrom. 2003;
38: 174.
54. Keire DA, Strauss E, Guo W, Noszal B, Rabenstein DL. Kinetic
and equilibria of thiol/disulfide interchange reactions of
selected biological thiols and related molecules with oxidized
glutathione. J. Org. Chem. 1992; 57: 123.
55. Millis KK, Weaver KH, Rabenstein DL. Oxidation/reduction
potential of glutathione. J. Org. Chem. 1993; 58: 4144.
56. Wei H, Qiu L, Divine KK, Ashbaugh MD, McIntyre LC,
Fernando Q, Gandolfi AJ. Toxicity and transport of three
synthesized mercury-thiol-complexes in isolated rabbit renal
proximal tubule suspensions. Drug Chem. Toxicol. 1999; 22: 323.
57. Ristoff E, Larsson A. Patients with genetic defects in the gamma-
glutamyl cycle. Chem.-Biol. Interact. 1998; 24: 111.
58. Mayatepek E. 5-Oxoprolinuria in patients with and without
defects in the gamma-glutamyl cycle. Eur. J. Pediatr. 1999;
158: 221.
59. Zalups RK. Molecular intertactions with mercury in the kidney.
Pharmacol. Rev. 2000; 52: 113.
66. Schafer F, Buettner GR. Redox environment of the cell as viewed
through the redox state of the glutathione disulfide/glutathione
couple. Free Radical Biol. Med. 2001; 30: 1191.
67. Goering PL. Lead-protein interactions as a basis for lead toxicity.
Neurotoxicology 1993; 14: 45.
68. Stohs SJ, Bagchi D. Oxidative mechanism in the toxicity of metal
ions. Free Radical Biol. Med. 1995; 18: 321.
69. Stohs SJ, Bagchi D, Hassoun E, Bagchi M. Oxidative mechanisms
in the toxicity of chromium and cadmium ions. J. Environ. Pathol.
Toxicol. Oncol. 2000; 19: 201.
70. Valko M, Morris H, Cronin MTD. Metals, toxicity and oxidative
stress. Curr. Med. Chem. 2005; 12: 1161.
71. Divine KK, Ayala-Fierro F, Barber DS, Carter DE. Glutathione,
albumin, cysteine and cys-gly effects on toxicity and
accumulation of mercuric chloride in LLC-PK1 cells. J. Toxicol.
Environ. Health A 1999; 57: 489.
72. Woods JS, Calas CA, Aicher LD, Robinson BH, Mailer C.
Stimulation of porphyrinogen oxidation by mercuric ion. I.
Evidence of free radical formation in the presence of thiols
and hydrogen peroxide. Mol. Pharmacol. 1990; 38: 253.
73. Miller DM, Woods JS. Redox activities of mercury-thiol
complexes: implications for mercury-induced porphyria and
toxicity. Chem.-Biol. Interact. 1993; 86: 23.
74. Yoshida M, Watanabe C, Kishimoyo M, Yasutake A, Satoh M,
Sawada M, Akama Yoshifumi. Behavioral changes in
metallothionein-null mice after the cessation of long-term low
level exposure to mercury vapor. Toxicol. Lett. 2006; 161: 210.
75. Hansen JM, Zhang H, Jones DP. Differential oxidation of
thioredoxin-1, thioredoxin-2, and glutathione by metal ions.
Free Radical Biol. Med. 2006; 40: 138.
76. Woods JS, Calas CA, Aicher LD. Stimulation of porphyrinogen
oxidation by mercuric ion. II. Promotion of oxidation from
the interaction of mercuric ion, glutathione and mitochondria-
generated hydrogen peroxide. Mol. Pharmacol. 1990; 38: 261.
77. Halbach S, Clarckson TW. Enzymatic oxidation of mercury
vapour by erythrocytes. Biochim. Biophys. Acta 1978; 523: 522.
78. Smith VC, Jones DP, Guenthner TM, Lash LH, Lautemburg BH.
Compartmentation of glutathione: implications for the study of
toxicity and disease. Toxicol. Appl. Pharmacol. 1996; 140: 1.
79. Wang Y, Fang J, Leonard SS, Krishna Rao MK. Cadmium
inhibits the electron transfer chain and induces reactive oxygen
species. Free Radical Biol. Med. 2004; 36: 1434.
80. Ganem B, Li Y-T, Henion JD. Detection of non-covalent receptor-
ligand complexes by mass spectrometry. J. Am. Chem. Soc. 1991;
113: 6294.
81. Rostom AA, Robinson CV. Disassembly of intact multiprotein
complexes in the gas phase. Curr. Opin. Struct. Biol. 1999; 9: 135.
82. Afonso C, Hathout Y, Fenselau C. Qualitative characterization
of biomolecular zinc complexes by collisionally induced
dissociation. J. Mass Spectrom. 2002; 37: 755.
60. Simmons-Willis TA, Koh AS, Clarkson TW, Ballatori N.
Transport of a neurotoxicant by molecular mimickry: the
methylmercury-l-cysteine complex is a substrate for human
L-type large neutral amino acid transporter (LAT) 1 and LAT2.
Biochem. J. 2002; 367: 239.
61. Ballatori N, Wang W, Liebermann MW. Accelerated methylmer-
cury elimination in ꢀ-glutamyl transpeptidase-deficient mice.
Am. J. Pathol. 1998; 152: 1049.
62. Hoffmeyer RE, Singh SP, Doonan CJ, Ross ARS, Hughes RJ,
Pickering IJ, George GN. Molecular mimicry in mercury
toxicology. Chem. Res. Toxicol. 2006; 19: 753.
63. Ferreira CS, Martins PS, Demicheli C, Brochu C, Ouellette M,
Frezard F. Thiol-induced reduction of antimony(V) into
antimony(III):
a comparative study with trypanothione,
cysteinyl-glycine, cysteine and glutathione. Biometals 2003; 16:
441.
83. Garcia JS, Schmidt de Magalhaes C, Zezzi Arruda MA. Trends
in metal-binding and metalloprotein analysis. Talanta 2006; 69: 1.
84. Mayumi Shindo, Kazuhiro Irie, Hiroyuki Fukuda, Hajime
Ohigashi. Analysis of the non-covalent interaction between
metal ions and the cysteine-rich domain of protein kinase C
eta by Electrospray ionization mass spectrometry. Bioorg. Med.
Chem. 2003; 11: 5075.
64. Enoiu M, Aberkane H, Salazar JF, Leroy P, Groffen J, Siest G,
Wellman M. Evidence for the pro-oxidant effect of gamma-
glutamyltranspeptidase-related enzyme. Free Radical Biol. Med.
2000; 29: 825.
65. Griffith OW, Tate SS. The apparent glutathione oxidase activity
of ꢀ-glutamyl transpeptidase. J. Biol. Chem. 1980; 255: 5011.
Copyright  2006 John Wiley & Sons, Ltd.
J. Mass Spectrom. 2006; 41: 1578–1593
DOI: 10.1002/jms