Modulation of nitric oxide-mediated metal release from
metallothionein by the redox state of glutathione
in vitro
Leila Khatai
1
, Walter Goessler
2
, Helena Lorencova
2
and Klaus Zangger
1
1
Institute of Chemistry, Organic and Bioorganic Chemistry, University of Graz, Austria;
2
Institute of Chemistry, Analytical Chemistry,
University of Graz, Austria
Metallothioneins (MTs) release bound metals when exposed
to nitric oxide. At inflammatory sites, both metallothionein
and inducible nitric oxide synthase (iNOS) are induced by
the same factors and the zinc released from metallothionein
by NO suppresses both the induction and activity of iNOS.
In a search for a possible modulatory mechanism of this
coexpression of counteracting proteins, we investigated the
role of the glutathione redox state in vitro because the oxi-
dation state of thiols is involved in the metal binding in Cd-S
or Zn-S clusters found in metallothioneins, and NO also
binds to reduced glutathione via S-nitrosation. Using a
variety of techniques, we found that NO and also ONOO
–
-
mediated metal release from purified MTs is suppressed by

2+
,Cd
2+
). Mammalian MTs bind seven
divalent metals in two separate domains [4]. Three metals
are bound in an M
3
Cys
9
cluster in the N-terminal b-domain,
while an M
4
Cys
11
four metal cluster is formed in the
C-terminal a-domain [4]. Of the four known mammalian
MT isoforms [2], the two best studied and most widely
occurring isoforms (1 and 2) are most abundant in
parenchymatous tissues, i.e. liver, kidney, pancreas and
intestines [5–7] but their occurrence and biosynthesis have
been documented in many tissues and cell types. The 3D
structures of MT1 [8] and MT2 [9–12] are very similar, but
there are various indications of increased flexibility and
metal mobility in the b-domaininMT-1[8].Thenaturally
bound metal zinc can be displaced by cadmium up to about
5 mol per mol protein by simple addition of Cd
2+
[13]
in vitro. Living animals fed a cadmium-rich diet produce
a mixed-metal MT with zinc bound preferentially in the

iNOS, inducible NO-synthase; MT, metallothionein; NO, nitric oxide;
SEC–ICPMS, size exclusion chromatography–inductively coupled
plasma mass spectrometry; SIN-HCl, 3-morpholinosydnoni-
mine.HCl.
(Received 26 February 2004, revised 6 April 2004,
accepted 14 April 2004)
Eur. J. Biochem. 271, 2408–2416 (2004) Ó FEBS 2004 doi:10.1111/j.1432-1033.2004.04160.x
Based on the preferred release of metal from the
b-domain of mouse MT1, where zinc is preferentially bound
in vivo, we suggested recently that MTs had anti-inflamma-
tory activity [31]. This activity relies on the suppression of
the expression and activity of inducible nitric oxide synthase
(iNOS) by zinc [32,33], released from MT under the
influence of nitric oxide (NO), and the scavenging of NO
through covalent binding to MTs [23] to form S-nitroso-
thiols. Such a role of MTs in the inflammatory response has
been corroborated by the significantly altered inflammatory
behavior during experimental autoimmune encephalomye-
litis [34] observed in MT deficient mice. In addition, it has
been reported that overexpression of MT reduces the
sensitivity of eukaryotic cells to oxidative injury [35] and the
cytotoxic effects of NO [29]. As both iNOS and MTs are
induced at inflammatory sites by the same compounds and
MT scavenges NO and suppresses its production, one starts
to wonder why they are both produced at inflammatory
sites but counteract each other. Therefore, we looked for a
possible regulatory mechanism for the interplay between
NO production and metal release from MTs in order to
understand this dual expression of opposing proteins. As
the metal is held in place by thiolate ligands [4] in MTs,

113
Cd-NMR spectroscopy as Zn cannot be studied by
regular NMR experiments. However, Cd
7
-MT1 is never
found in natural sources and differences in metal-binding
constants between Cd and Zn might prevent the inter-
pretation of in vivo processes with data obtained on an
artificially cadmium-enriched protein. Therefore, we have
limited the present study to Zn
7
-MT2 and Cd
5
Zn
2
-MT2,
which have been isolated from natural sources.
In addition to nitric oxide, peroxynitrite (ONOO
–
)may
also play a significant role in the metal release from MTs, as
it has been suggested that the decomposition of peroxy-
nitrite at physiological pH constitutes the actual component
of NO cytotoxicity [39]. A widespread signal transduction
mechanism for NO involved in, e.g. platelet aggregation,
blood pressure control and neurotransmission functions
via stimulation of guanylyl cyclase [40]. In contrast to NO,
glutathione-dependent bioactivation of peroxynitrite is
involved in enzyme stimulation and this points again at a
possible key role of glutathione in the NO and/or ONOO

per protein monomer according to the procedure des-
cribed by Vas
ˇ
ak [15], but found no differences in its
behavior to the unpurified commercially available product.
All other chemicals were purchased from Sigma (Vienna,
Austria) at the highest purity available. Due to problems
associated with the use of organic buffers in inductively
coupled plasma mass spectrometry (ICPMS) instruments
and protonated buffers for NMR, we used aqueous
phosphate buffers for all experiments (see below). To
evaluate a possible influence of phosphate ions on MT2
during these experiments, the CD experiments were also
performed in 20 m
M
Hepes buffer, but showed the same
results.
CD spectroscopy
The complete absence of aromatic amino acids in
metallothioneins allows the use of UV and CD spectros-
copy to observe the cadmium-thiolate charge transfer
transition, which occurs around 250 nm [42]. This region
is usually completely masked by aromatic groups. CD
spectra were recorded on a Jasco J-715 spectropolarimeter
andanalyzedusingtheprogram
CDSCAN
. For each
wavelength scan, the average was taken from 10 accumu-
lations with the following parameters: step resolution,
0.2 nm; speed, 50 nmÆmin

time scans were obtained by monitoring the CD at the
maximum of the cadmium-thiolate charge transfer band
at 260 nm for 20 min after mixing the components.
Ó FEBS 2004 NO-mediated metal release from metallothionein (Eur. J. Biochem. 271) 2409
SEC–ICPMS
ICPMS enables the determination of a variety of elements
in solution. In order to differentiate between protein-bound
and free metal, a preceding separation of protein and
unbound metal by size exclusion chromatography (SEC) is
necessary. Instead of performing these two steps separately,
thecouplingofSECandICPMSoffersaveryelegant
alternative [43,44]. For our studies, a Pharmacia Superdex
75 PC 3.2/30 gel filtration column was connected to an
Agilent HP 1100 ChemStation SEC system (Agilent,
Waldbronn, Germany) equipped with a UV monitor set
to 220 nm. The outlet of the UV-detector was connected
directly via a PEEK capillary (i.d. 0.12 mm, length 90 cm)
to the l-flow PFA-100 nebulizer (CPI International, Santa
Rosa, USA) of the Agilent 7500c ICPMS system. The
isotopes
64,66
Zn and
111,114
Cd were monitored. All meas-
urements were performed at least twice and the averages
were taken over both isotopes of zinc and cadmium,
respectively. A 20 m
M
aqueous ammonium phosphate
buffer, pH 6.5 was used as eluent at a flow rate of

injection onto the gel filtration column.
NMR spectroscopy
Series of two-dimensional TOCSY [45] NMR spectra were
recorded on a Varian Unity INOVA 600 MHz NMR
spectrometer at 25 °C. The water signal was suppressed
with the WATERGATE sequence [46]. For each of the 256
increments, 2048 complex data points were recorded. The
data were multiplied with a 60° phase-shifted, squared sine-
bell window function in both dimensions prior to Fourier
transformation. The total experimental time of one 2D
spectrum was 12 h. Samples consisted of 2.5 mg of Zn
7
-
MT2 or Cd
5
Zn
2
-MT2 in 0.5mL of 20m
M
potassium
phosphate buffer pH 6.5 and 50 lLD
2
O. A stock solution
of 50 m
M
DEA/NO was added directly to the NMR
samples to give final concentrations of 0.2, 0.5, 1 and 3 m
M
DEA/NO. After each addition, the solution was equili-
brated for at least 20 min prior to the start of the NMR

-MT2 was exposed to NO for 20 min by adding
DEA/NO at a final concentration of 1 m
M
. For reduced
and oxidized glutathione, concentrations of 1 m
M
were
used. CD spectra of the range between 230 and 300 nm are
shown in Fig. 1A. The Cd-S charge transfer band at 260 nm
is reduced clearly after the addition of NO, indicating the
breaking of cadmium-cysteine bonds and therefore release
of cadmium. The presence of GSH reduces the metal release
almost completely, while GSSG even slightly increased the
cadmium release by nitric oxide. The decay occurs in the
first 10 min after the addition of NO as observed by
monitoring the CD at the maximum of the charge transfer
band at 260 nm (Fig. 1B). The lower molar ellipticity at
time 0 in the GSSG/NO treated sample derives from partial
metal release during the period from mixing the solutions
until the start of the data acquisition. While with the CD
measurements nothing can be said about the faith of zinc
bound in Cd
5
Zn
2
-MT2 after the addition of NO, the
amount of cadmium is reduced at this rather extreme NO
Fig. 1. Wavelength and time scans of the CD of MT2 in the presence
and absence of NO and GSH/GSSG. CD spectra of 100 l
M

sample is applied to a gel filtration column, which separates
free from protein-bound metal and subsequently both zinc
and cadmium levels are determined by ICPMS. Stock
solutions of 2 m
M
DEA/NO, 10 m
M
GSH and 5 m
M
GSSG were added to samples of 20 l
M
Cd
5
Zn
2
-MT2 to
give final ratios as indicated at the bottom of Fig. 2. The
normalized amounts of zinc and cadmium in the MT2
fraction, taking into consideration the dilution effects by
adding stock solutions of GSH, GSSG and DEA/NO
(Fig. 2) clearly show that the release of both cadmium and
zinc by NO is suppressed completely by GSH, but not
GSSG. As already suggested in our previous paper [31],
zinc is more readily released than cadmium. Rather high
concentrations of NO are needed to observe significantly
reduced cadmium levels in MT2, which corroborates the
role of MTs in heavy metal detoxification as a result of
rather tight binding of cadmium to MTs [18,19]. The
maximum number of metals released from Cd
5

thione [39]. To elucidate the possible role of peroxynitrite in
the metal release from MTs in the presence of reduced and
oxidized glutathione we carried out SEC–ICPMS measure-
ments on a series of solutions containing a mixture of
Cd
5
Zn
2
-MT2 (20 l
M
stock solution), the peroxynitrite
donor SIN-HCl (2 m
M
stock solution), a freshly prepared
peroxynitrite solution (2 m
M
stock solution) and either
GSH (10 m
M
stock solution) or GSSG (10 m
M
stock
solution) at the ratios shown in Fig. 4. As can be seen, the
metal release by both SIN-HCl and ONOO
–
is not as
pronounced as for NO itself, leading to a maximum of
about 1.2 Zn at far from physiological NO/MT ratios of
100 : 1 and only insignificant amounts of cadmium being
released at the highest ONOO

protein fraction of the SEC–ICPMS chromatograms with relative error
bars in the presence of NO, GSH and/or GSSG. A solution of 20 l
M
Cd
5
Zn
2
-MT2 was diluted with stock solution of 2 m
M
DEA/NO,
10 m
M
GSH and 5 m
M
GSSG to give ratios of these compounds as
indicated at the bottom.
Ó FEBS 2004 NO-mediated metal release from metallothionein (Eur. J. Biochem. 271) 2411
only well-resolved peaks were integrated and their signal
intensities normalized to the intensity in the absence of NO
(I
0
). Representative NO-concentration dependences for all
well-resolved signals from the a- (22 peaks) and b-domain
(31 peaks) were averaged and are shown in Fig. 5. The
reductions in proton signal intensities reflect the increase of
dynamic processes when metal is released and/or the
conformational variety in the disulfide bridged MT2 formed
after NO treatment as described [31] and so it can be used
indirectly to follow metal binding stochiometries. The
addition of NO at these high concentrations leads to signal

difference in the amount of metal released by NO in the two
domains of MT3 [49]. Zinc from the b-domain was set free
much easier than from the a-domain. Thus, the already
observed distinctive metal mobilities in b-domains of
MT isoforms 1, 2 and 3, which follow the order
MT3 > MT1 > MT2 [8,50,51] are mirrored in the metal
release upon NO exposure.
Discussion
The presented results show clearly that the metal release
from MT2 by nitric oxide and peroxynitrite is suppressed
by reduced but not oxidized glutathione. Due to different
requirements of sample concentrations in the presented
experiments, the interaction of MT2, NO, ONOO
–
and
GSH has been established for ratios ranging from
1:0.3:1.4upto1:10:100(MT:NO/ONOO
–
:GSH)
with MT2 being between 20 and 600 l
M
. The reason for
NO protection by glutathione could be attributed to its
faster reaction with NO or the reported binding of GSH in
the b-domain of metallothionein [52,53] and thus the
blocking of certain nitrosation sites. Surprisingly, we did
not observe any changes in the TOCSY NMR spectra upon
the addition of GSH (data not shown), which is indicative of
no specific binding under the conditions (buffer system and
pH) used here. Still, the suppression of the NO–MT2

inflammation [58,59] but we are not aware of any report of
GSH concentrations low enough to enable metal release
from MT2 upon the exposure to nitric oxide or peroxy-
nitrite. However, a number of reports have been published
demonstrating the physiological significance of the NO–MT
interaction and in particular the metal release in vivo.Using
a fluorescent MT2 fusion protein, a conformational change
in MT2, indicative of metal- release, has been observed by
Pearce et al. [28] after the administration of NO or NO-
stimulating factors in endothelial cells. The metal release
itself has been studied in cultured epithelial cells [26].
Metallothionein has also been shown to protect eukaryotic
cells from the cytotoxic and DNA-damaging effects of nitric
oxide [29]. So, while the binding of NO to GSH in vivo does
not obviously prevent NO from interacting with MT2,
we have shown that in vitro it suppresses the metal release
from metallothioneins. This points to an hitherto unknown
mechanism or compound(s) being involved in this inter-
action in living cells and information about this additional
factor is needed in order to perform physiologically relevant
future in vitro studies and in the interpretation of
results obtained from in vivo experiments on the NO–MT
interaction.
As predicted earlier [31], zinc is more readily released
from MTs than cadmium, which is probably a combination
of tighter binding of cadmium than zinc in metallothioneins
and the preference of zinc in the more flexible b-domain. In
addition to the already described differences in flexibility of
the b-domain in MT isoforms 1 and 2 [8], we found that the
domain specific distinctions upon NO exposure are less

a-domain (bottom) exposed to NO in the absence or presence of 5 m
M
GSH. The average intensities of all intense, well-resolved peaks in the
2D TOCSY spectra (22 peaks from the b-domain and 31 from the
a-domain) with 0 and 3 m
M
DEA/NO were used.
Ó FEBS 2004 NO-mediated metal release from metallothionein (Eur. J. Biochem. 271) 2413
significant in MT2 unlike previously found for mouse Cd
7
-
MT1 [31].
In conclusion, we have shown that reduced but not
oxidized glutathione suppresses the NO and ONOO
–
-
mediated metal release from metallothionein in vitro and
that zinc is indeed more readily released under these
conditions as suggested earlier [31]. The millimolar concen-
trations of GSH present in mammalian cells should thus
eliminate any nitric oxide or peroxynitrite mediated metal
release from MTs. However, as such an interaction has been
found in vivo, an unknown mechanism or compound must
also be involved in this interaction. Therefore, we believe
that results from both in vivo and in vitro studies on the
NO–MT interaction should be interpreted with caution for
as long as this discrepancy has not been resolved.
Acknowledgements
This work has been supported by the Austrian Science Foundation
(Project No. P15289 to K. Z.). We would like to thank Monika Oberer

by homonuclear and heteronuclear NMR spectroscopy. Protein
Sci. 8, 2630–2638.
9. Arseniev, A., Schultze, P., Wo
¨
rgo
¨
tter,E.,Braun,W.,Wagner,G.,
Vasak, M., Ka
¨
gi,J.H.&Wu
¨
thrich, K. (1988) Three-dimensional
structure of rabbit liver [Cd7]metallothionein-2a in aqueous
solution determined by nuclear magnetic resonance. J. Mol. Biol.
201, 637–657.
10. Messerle, B.A., Scha
¨
ffer,A.,Vasak,M.,Ka
¨
gi,J.H.&Wu
¨
thrich,
K. (1990) Three-dimensional structure of human [113Cd7]metal-
lothionein-2 in solution determined by nuclear magnetic resonance
spectroscopy. J. Mol. Biol. 214, 765–779.
11. Robbins, A.H., McRee, D.E., Williamson, M., Collett, S.A.,
Xuong,N.H.,Furey,W.F.,Wang,B.C.&Stout,C.D.(1991)
Refined crystal structure of Cd, Zn metallothionein at 2.0 A
resolution. J. Mol. Biol. 221, 1269–1293.
12. Schultze, P., Wo

Koropatnick, J., Lazo, J.S. & Waalkes, M.P. (1994) Role of
metallothionein in carcinogenesis. Toxicol. Appl. Pharmacol. 126,
1–5.
20. Sato, M. & Bremner, I. (1993) Oxygen free radicals and
metallothionein. Free Radic. Biol. Med. 14, 325–337.
21. Thornalley, P.J. & Vasak, M. (1985) Possible role for
metallothionein in protection against radiation-induced oxidative
stress. Kinetics and mechanism of its reaction with superoxide and
hydroxyl radicals. Biochim. Biophys. Acta. 827, 36–44.
Fig. 6. Changes in proton NMR peak intensity as a function of DEA/
NO concentration are shown for one representative signal of the b-do-
main (Ala8 CH
3
) and one from the a-domain (Ala53 CH
3
). NO-con-
centration dependent intensities of: j,Zn
7
-MT2; m,Cd
5
Zn
2
-MT2 and
s,Zn
7
-MT2 in the presence of up to 5 m
M
GSH. Standard deviations
were extracted from the observed signal to noise ratio.
2414 L. Khatai et al.(Eur. J. Biochem. 271) Ó FEBS 2004

green fluorescent fusion protein. Proc. Natl Acad. Sci. USA. 97,
477–482.
29. Schwarz, M.A., Lazo, J.S., Yalowich, J.C., Allen, W.P.,
Whitmore, M., Bergonia, H.A., Tzeng, E., Billiar, T.R., Robbins,
P.D.,Lancaster,J.R.Jr,et al. (1995) Metallothionein protects
against the cytotoxic and DNA-damaging effects of nitric oxide.
Proc.NatlAcad.Sci.USA. 92, 4452–4456.
30. St Croix, C.M., Wasserloos, K.J., Dineley, K.E., Reynolds, I.J.,
Levitan, E.S. & Pitt, B.R. (2002) Nitric oxide-induced
changes in intracellular zinc homeostasis are mediated by
metallothionein/thionein. Am.J.Physiol.LungCellMol.Physiol.
282, L185–L192.
31. Zangger, K., O
¨
z,G.,Haslinger,E.,Kunert,O.&Armitage,I.M.
(2001) Nitric oxide selectively releases metals from the amino-
terminal domain of metallothioneins: potential role at inflam-
matory sites. Faseb J. 15, 1303–1305.
32. Persechini, A., McMillan, K. & Masters, B.S. (1995) Inhibition of
nitric oxide synthase activity by Zn2+ ion. Biochemistry 34,
15091–15095.
33. Abou-Mohamed, G., Papapetropoulos, A., Catravas, J.D. &
Caldwell, R.W. (1998) Zn2+ inhibits nitric oxide formation in
response to lipopolysaccharides: implication in its anti-inflamma-
tory activity. Eur. J. Pharmacol. 341, 265–272.
34. Penkowa, M., Espejo, C., Martinez-Caceres, E.M., Poulsen, C.B.,
Montalban, X. & Hidalgo, J. (2001) Altered inflammatory
response and increased neurodegeneration in metallothionein
I+II deficient mice during experimental autoimmune encephalo-
myelitis. J. Neuroimmunol. 119, 248–260.

412–417.
44. Lobinski, R., Chassaigne, H. & Szpunar, J. (1998) Analysis
for metallothioneins using coupled techniques. Talanta 46,
271–289.
45. Braunschweiler, L. & Ernst, R.R. (1983) Coherence transfer by
isotropic mixing: application to proton correlation spectroscopy.
J. Magn. Reson. 53, 521–528.
46. Piotto, M., Saudek, V. & Sklenar, V. (1992) Gradient-tailored
excitation for single-quantum NMR spectroscopy of aqueous
solutions. J. Biomol. NMR. 2, 661–665.
47. Otvos,J.D.,Liu,X.,Li,H.,Shen,G.&Basti,M.(1993)Dynamic
aspects of metallothionein structure. In Metallothionein III
(Suzuki,K.T.,Imura,N.&Kimura,M.,eds),pp.57–74.
Birkha
¨
user-Verlag, Basel, Switzerland.
48. Palumaa, P., Mackay, E.A. & Vasak, M. (1992) Nonoxidative
cadmium-dependent dimerization of Cd7-metallothionein from
rabbit liver. Biochemistry 31, 2181–2186.
49.Chen,Y.,Irie,Y.,Keung,W.M.&Maret,W.(2002)
S-nitrosothiols react preferentially with zinc thiolate clusters of
metallothionein III through transnitrosation. Biochemistry 41,
8360–8367.
50. Hasler, D.W., Jensen, L.T., Zerbe, O., Winge, D.R. & Vasak, M.
(2000) Effect of the two conserved prolines of human growth
inhibitory factor (metallothionein-3) on its biological activity and
structure fluctuation: comparison with a mutant protein.
Biochemistry 39, 14567–14575.
51. O
¨

mation: therapeutic approaches. Free Radic. Biol. Med. 28, 1405–
1420.
2416 L. Khatai et al.(Eur. J. Biochem. 271) Ó FEBS 2004