Evidence for gliotoxin-glutathione conjugate adducts

Article abstract of DOI:10.1016/S0960-894X(00)00696-X

The equilibrium constant for the gliotoxin/glutathione pair was found to be 1200±100M^-1 at pH 7.0 at 25°C. Under conditions where the reaction was quenched rapidly with the addition of acid, gliotoxin-glutathione conjugate adducts were detected

Full text of DOI:10.1016/S0960-894X(00)00696-X

Bioorganic & Medicinal Chemistry Letters 11 (2001) 483±485
Evidence for Gliotoxin±Glutathione Conjugate Adducts
Paul H. Bernardo,a Christina L. L. Chai,a,* Geo€rey J. Deeble,a Xue-Ming Liua
and Paul Waringb
aDepartment of Chemistry, The Faculties, Australian National University, ACT 0200, Australia
bDivision of Cell Biology and Immunology, John Curtin School of Medical Research, Australian National University, ACT 0200, Australia
Received 11 October 2000; accepted 3 December 2000
AbstractÐThe equilibrium constant for the gliotoxin/glutathione pair was found to be 1200Æ100 MÀ1 at pH 7.0 at 25 ꢀC. Under
conditions where the reaction was quenched rapidly with the addition of acid, gliotoxin±glutathione conjugate adducts were
detected. # 2001 Elsevier Science Ltd. All rights reserved.
Gliotoxin (1) is a biologically active cyclic disul®de that
displays a wide range of biological e€ects including
antiviral, antibacterial, and immunosuppressive proper-
ties.1 Implicated in the mechanism of action of gliotoxin
is the interaction of the disul®de linkage with sulfur
nucleophiles in a thiol-disul®de exchange mechanism.2
In view of the abundance of intracellular glutathione
(GSH), it is likely that mixed disul®des of glutathione
with gliotoxin are potential intermediates in cell systems
treated with gliotoxin. Despite this, no evidence for the
formation of glutathione±gliotoxin adducts has been
reported. In this communication, we report our
attempts to isolate and characterize these adducts using
a combination of HPLC, NMR, and MS techniques.
equilibrium constants, a mixture of gliotoxin and a
redox bu€er (glutathione GSH/oxidized glutathione
GSSG) was equilibrated under strictly degassed condi-
tions. Samples were quenched with acid (to pH 2) prior
to analysis by reverse-phase HPLC analysis. In the
typical solvent system used to monitor the progress of
reaction (see Experimental), the retention times of glio-
toxin (1) (17 min) and reduced gliotoxin (2) (15.8 min)
were suciently di€erent to enable accurate integration
of the areas. To ensure that the position of equilibrium
had been reached, the reaction mixture was monitored
after both 1 and 5 h and in all cases, the values obtained
were constant. In addition, perturbation of equilibrium
by the addition of the corresponding disul®de, GSSG,
gave similar Keq values. The equilibrium constant at pH
7.0 was measured as 1200Æ100 MÀ1 for the gliotoxin/
glutathione pair, giving rise to an Eo value of À0.17 V
for the reduction of gliotoxin to reduced gliotoxin. As a
comparison, similar equilibrium studies were carried out
using gliotoxin and cysteine and the measured Keq
value was 400Æ4 MÀ1. From these two Keq values, the
Keq value for glutathione with cystine is calculated to
be 3.0, which is in good agreement with the literature
values.3
Our preliminary studies show that the rate at which
equilibrium between gliotoxin and glutathione is estab-
lished is dependent on a number of factors including the
concentrations of glutathione and pH of the media. For
example, at high concentrations of glutathione, the
equilibrium is rapidly established (typically within
30 min). Similarly equilibration is more rapid at high
pH values of the media. In these cases, there was no
evidence for the formation of a gliotoxin±glutathione
adduct by HPLC. For a typical measurement of
To optimize the observation of intermediate species,
samples of gliotoxin and glutathione were incubated at
pH 7.0 and the aliquots of the reaction were quenched
at various time intervals with tri¯uoroacetic acid
(TFA)/water. Analysis of the samples by HPLC showed
that at fast quenching times (less than 30 s), new peaks
at 10.8, 11.8, and 13.6 min were observed. These peaks
were greatly diminished at higher quenching times, and
provide evidence for the transient lifetime of these species.
*Corresponding author. Fax: +61-2-6125-0760; e-mail: christina.chai@
anu.edu.au
0960-894X/01/$ - see front matter # 2001 Elsevier Science Ltd. All rights reserved.
PII: S0960-894X(00)00696-X

484
P. H. Bernardo et al. / Bioorg. Med. Chem. Lett. 11 (2001) 483±485
In order to determine the identities of the intermediate
species, rapidly quenched samples were analyzed by HPLC
and fractions corresponding to the new peaks were col-
lected, lyophilized and re-analyzed. The peak at 10.8 min
corresponded to a stable compound with a molecular for-
mula of 938 Da as determined by ES-MS analysis.4 This
is consistent with the formulation of a double mixed
disul®de (DMD) species with the structure 3.
Figure 1. Possible routes for the interaction of gliotoxin with glu-
tathione.
NMRanalysis in D O in the presence of 0.1% deuter-
2
exchange reactions than the other (MD 1, 13.6 min
fraction). The reason for this is not clear and may be
related to increased steric congestion of the disul®de
linkage in one mono mixed disul®de (presumably 5) as
compared to the other. Thus intramolecular thiol di-
sul®de exchange reaction to form 4 and intramolecular
cyclization to form gliotoxin will relieve some of the
steric strain involved. It has also been established that
the rate constant of thiol-disul®de exchange reactions
correlate with the fraction of reactive thiolate ions pre-
sent in solution. This in turn depends on the pKa of the
reacting thiol as well as the pH of the solution.8 Thus it
is also conceivable that the propensity of MD2 to con-
vert to MD1 may be related to the relative acidities of
the thiol groups.
ated TFA is complicated by a number of overlapping
signals but resonances attributed to the gliotoxin and
glutathione portion of the structure are clearly evident
from 1-D and 2-D NMRanalysis.
In contrast, the peaks corresponding to the retention
times 11.8 and 13.6 min are much less stable. In the
HPLC analysis of the collected, lyophilized fraction at
11.8 min, interconversion to the species at 13.6 min as
well as to gliotoxin (17 min) is observed. This provides
indirect evidence that the species at 11.8 and 13.6 min
are mono mixed disul®des 4 and 5. The mono mixed
disul®de corresponding to the retention time 13.6 min is
evidently the more stable of the two mono mixed di-
sul®des. The fraction corresponding to this was col-
lected, lyophilized and when analyzed by ES-MS, a
molecular ion at 665 was observed.5 This is consistent with
the formulation of C23H31N5O12S3, that is, the molecular
ion of the mono mixed disul®de+O2,6 presumably arising
from aerial oxidation in the isolation process.
The double mixed disul®de species isolated can be envi-
saged as arising from the mono mixed disul®des via
aerial oxidation with glutathione or via interaction of
the thiol group of mono mixed disul®des with oxidized
glutathione. As the double mixed disul®de species is not
observed when quenching of the gliotoxin/glutathione
reaction mixture is slow, this species must be reduced to
the dithiol in the presence of excess glutathione. The
equilibrium constant of the gliotoxin/glutathione pair
indicates that reduction of the disul®de linkage of glio-
toxin is strongly favoured and this implies that at
physiological pH, the toxin is capable of depleting cel-
lular glutathione. Related to this, reduced gliotoxin has
been observed in treated cells.9
These studies show that the remaining thiol group in
each of the mono mixed disul®des is clearly very reac-
tive and is prone to intramolecular SN2 reactions or to
aerial oxidation. Attempts were made to derivatize the
remaining thiol group of 4 and 5 via oxidation with
hydrogen peroxide and by alkylation with iodoacetic acid.
The former method gave a complex mixture of com-
pounds whereas alkylation at pH 7.0 was slow and does
not compete e€ectively with intramolecular reactions.7
Gliotoxin, glutathione (GSH), and oxidized glutathione
(GSSG) were purchased from Sigma. HPLC analyses
were carried out using a Beckmann instrument consist-
ing of a 126 Dual Pump Solvent Delivery System ®tted
with a UV detector. A wavelength of 260 nM was used
unless otherwise speci®ed. A 0.46 cmÂ15 cm Beckman
Ultrasphere ODS 5 micron reverse-phase column was
used. Studies were carried out using 5% acetonitrile for
the ®rst 5 min, followed by a 5±40% gradient of aceto-
nitrile in water over 15 min in the presence of 0.1% TFA
and with argon bubbling through the solvent reservoirs
at all times. For the preparation of the mixed disul®des,
samples were isolated using an auto-fraction collector
and were kept on ice or dry-ice before analysis. Elec-
trospray MS was recorded on a Fisons VG Quattro II
spectrometer equipped with a Hewlett Packard 1090
LCS.
The studies above show that glutathione interacts with the
cyclic disul®de moiety of gliotoxin via a thiol disul®de
exchange mechanism (Fig. 1). Due to the unsymmetrical
nature of the disul®de, two mono mixed disul®des are
formed. The two mono mixed disul®des can in principle
undergo intramolecular thiol disul®de exchange as well
as intramolecular cyclization to gliotoxin. From our cur-
rent studies, it appears that one mono mixed disul®de
species (MD 2, 11.8 min fraction) is more reactive
towards inter- and intramolecular thiol-disul®de

P. H. Bernardo et al. / Bioorg. Med. Chem. Lett. 11 (2001) 483±485
485
Standards of gliotoxin and reduced gliotoxin as well as
reduced and oxidized glutathione were sampled on the
HPLC in order to determine their respective retention
times. Concentrations of the glutathione solution were
con®rmed by the addition of Ellman's Reagent followed
by UV±vis measurements. Standard curves were
obtained independently for gliotoxin and reduced glio-
toxin on the HPLC
quenched almost immediately (10±30 s) by the addition
of 600 mL of a 10% TFA/H2O solution. The quenched
mixture was then lyophilized, then redissolved in
300 mL of 1.67% TFA/H2O solution, followed by
HPLC separation.
Acknowledgements
The authors gratefully acknowledge the Australian
Research Council for ®nancial support and the China
Scholarship Council for a scholarship for X.-M. Liu.
A typical procedure for the determination of equilibrium
constants
To 2.8 mL of phosphate bu€er at pH 7 in a Reacti-vial
were added 100 mL of a 30 mM solution of GSH and
100 mL of a 30 mM solution of GSSG. This GSH/GSSG
redox bu€er was mixed by gentle inversion. To initiate
the reaction, 100 mL of a 3 mM solution of gliotoxin was
added and the resulting solution was mixed by inversion.
At 1 and 5 h later, a 300 mL aliquot was removed from
the reaction mixture under nitrogen and placed in a
HPLC vial containing 1 mL tri¯uoroacetic acid (TFA).
The vial was then sealed under a nitrogen atmosphere
and the aliquot was immediately analyzed by HPLC.
References and Notes
1. (a) Taylor, A. In Microbial Toxins; Kadis, S., Ciegler, A.,
Ajl, S. J., Eds.; Academic: New York, 1971; Vol. 7, pp 337±376.
(b) Waring, P.; Beaver, J. P. J. Gen. Pharmacol. 1996, 27, 1311.
(c) Waring, P.; Eichner, R.; Mullbacher, A. Med. Res. Rev.
È
1988, 8, 499. (d) Chai, C. L. L.; Waring, P. Redox Reports 2000,
in press.
2. (a) Jordan, T. W.; Cordiner, S. J. TIPS 1987, 8, 144. (b)
Hurne, A. M.; Chai, C. L. L.; Waring, P. J. Biol. Chem. 2000,
275, 25202.
Gliotoxin and reduced gliotoxin concentrations were
determined by comparison of peak areas to standard
curves. Concentrations of GSH and GSSG were treated
as constant since both compounds were in large excess,
and the systematic errors were within 5%.
3. Gorin, G.; Doughty, G. Arch. Biochem. Biophys. 1968, 126,
547.
4. C33H46N8O16S4 (ES+, 50 V, m/z) 940.9 (10%), 939.8
(15%), 938.8 (35%), 633.9 (7%), 470 (70%), 338.8 (67%), 308
(81%), 262.9 (56%), 86 (100%)
5. C23H31N5O12S3 (ES+, 50 V, m/z) 668 (22%), 666.9 (32%),
665.9 (100%), 633.9 (26%), 585.9 (15%), 584.9 (25%), 583.9
(90%).
Similar procedures were used with di€erent concentra-
tions of GSH/GSSG, as well as at di€erent pH's.
6. The oxidation of a thiol group to the sul®nic acid is not
common though this has been observed previously in alkaline
media: Berger, H. Rec. Trav. Chim. 1963, 82, 773.
7. For successful S-alkylation, the thiolate ion must ®rst be
generated. Our model studies with thiol piperazinediones in
organic solvents have shown that S-alkylation is dicult in the
absence of base.
8. Whitesides, G. M.; Lilburn, J. E.; Szajewski, R. P. J. Org.
Chem. 1977, 42, 332.
9. Waring, P.; Newcombe, N.; Edel, M.; Lin, Q. H.; Jiang, H.;
Typical procedure for the preparation of mixed disul®des
Five-hundred microlitres of a 12.9 mM gliotoxin/aceto-
nitrile solution were added to 5 mL of degassed phos-
phate bu€er (0.05 M, pH 7.0) solution under argon at
30 ꢀC. The solution was degassed by bubbling the reac-
tion mixture with Argon for 15 min. This was then fol-
lowed by the addition of 500 mL of a 193.5 mM solution
of glutathione in degassed bu€er. The reaction was
Sjaarda, A.; Piva, T.; Mullbacher, A. Toxicon 1994, 32, 491.
È