REDOX MODULATION OF CELL SURFACE PROTEIN THIOLS IN U937
LYMPHOMA CELLS: THE ROLE OF ?-GLUTAMYL TRANSPEPTIDASE-
DEPENDENT H2O2PRODUCTION AND S-THIOLATION
SILVIA DOMINICI,* MELISSA VALENTINI,* EMILIA MAELLARO,* BARBARA DEL BELLO,* ALDO PAOLICCHI,†
EVELINA LORENZINI,†ROBERTO TONGIANI,†MARIO COMPORTI,* and ALFONSO POMPELLA*
*Institute of General Pathology, University of Siena, Siena, Italy; and†Department of Experimental Pathology,
University of Pisa, Pisa, Italy
(Received 16 February 1999; Accepted 17 May 1999)
Abstract—The expression of gamma-glutamyl transpeptidase (GGT), a plasma membrane ectoenzyme involved in the
metabolism of extracellular reduced glutathione (GSH), is a marker of neoplastic progression in several experimental
models, and occurs in a number of human malignant neoplasms and their metastases. Because it favors the supply of
precursors for the synthesis of GSH, GGT expression has been interpreted as a member in cellular antioxidant defense
systems. However, thiol metabolites generated at the cell surface during GGT activity can induce prooxidant reactions,
leading to production of free radical oxidant species. The present study was designed to characterize the prooxidant
reactions occurring during GGT ectoactivity, and their possible effects on the thiol redox status of proteins of the cell
surface. Results indicate that: (i) in U937 cells, expressing significant amounts of membrane-bound GGT, GGT-
mediated metabolism of GSH is coupled with the extracellular production of hydrogen peroxide; (ii) GGT activity also
results in decreased levels of protein thiols at the cell surface; (iii) GGT-dependent decrease in protein thiols is due to
sulfhydryl oxidation and protein S-thiolation reactions; and (iv) GGT irreversible inhibition by acivicin is sufficient to
produce an increase of protein thiols at the cell surface. Membrane receptors and transcription factors have been shown
to possess critical thiols involved in the transduction of proliferative signals. Furthermore, it was suggested that
S-thiolation of cellular proteins may represent a mechanism for protection of vulnerable thiols against irreversible
damage by prooxidant agents. Thus, the findings reported here provide additional explanations for the envisaged role
played by membrane-bound GGT activity in the proliferative attitude of malignant cells and their resistance to
prooxidant drugs and radiation therapy.© 1999 Elsevier Science Inc.
Keywords—?-Glutamyl transpeptidase, Glutathione, Hydrogen peroxide, Protein thiols, Protein S-thiolation, Free
Adequate intracellular levels of the nucleophilic tripep-
tide GSH are crucial for the functioning of important
cellular systems operating in the defense against prooxi-
dant agents and for detoxification of electrophilic cyto-
toxic drugs. Despite the fact that uptake of intact GSH
does occur in small intestine and kidney [1,2], in most
cell types, extracellular GSH cannot cross plasma mem-
brane as such; thus, its intracellular levels depend on a
balance between its consumption and its de novo syn-
thesis; the latter in turn depends on an adequate supply of
precursor amino acids. In a number of cell types, such a
supply is warranted by ?-glutamyl transpeptidase (GGT;
EC 22.214.171.124), an ectoenzyme with the active site oriented
toward the outer cell surface, which is capable of cleav-
ing extracellular GSH. Precursor amino acids are thus
formed, which can cross the plasma membrane and are
re-utilized for intracellular GSH synthesis . Further-
more, because it has been documented that a continuous
efflux of GSH occurs from a number of cell types
through specific out-transporters , it appears that a
major function of GGT ectoactivity is to expedite the
salvage of extracellular GSH, which would otherwise be
lost from the cell. ?-Glutamyl transpeptidase thus ap-
Address correspondence to: Dr. A. Pompella, Institute of General
Pathology, Via Aldo Moro - I-53100 Siena, Italy; Tel: ?390 (577)
234-004; Fax: ?390 (577) 234-009; E-Mail: email@example.com.
Free Radical Biology & Medicine, Vol. 27, Nos. 5/6, pp. 623–635, 1999
Copyright © 1999 Elsevier Science Inc.
Printed in the USA. All rights reserved
0891-5849/99/$–see front matter
pears to participate in a “GSH cycling” in the plasma
The expression of significant levels of GGT has been
reported to occur in a number of human malignant neo-
plasms, such as ovary [6,7], colon [8,9], lung , liver
, sarcoma , melanoma , and leukemias [14,
15]; in many cases, levels of GGT detectable in metas-
tases are higher than in corresponding primitive localiza-
tions. In a series of 60 human tumor cell lines, GGT was
found to be significantly expressed in 70% of cases .
In addition, GGT expression is known to be a marker of
neoplastic progression in several experimental models,
such as rodent skin and liver chemical carcinogenesis
. Transfection of epithelial cells with the oncogene
ras, while resulting in the appearance of metastatic be-
havior, is also accompanied by expression of GGT [18,
19]. Interestingly, in melanoma clones, the degree of
GGT expression was found to be proportional to the
invasive and migrating abilities . Altogether, this and
other evidence concur in highlighting GGT activity as a
factor in human malignancy.
Because GGT plays a crucial role in the cellular
supply of GSH, thus favoring in tumor cells the appear-
ance of resistance against electrophilic chemotherapeu-
tics, GGT has often been interpreted as a member in the
cellular antioxidant enzyme systems [5,6,20]. However,
the significance of GGT in the cellular redox equilibrium
was made more complex by recent studies suggesting
that oxidant compounds (superoxide anion, hydrogen
peroxide, thiyl radicals) might be produced during GGT-
mediated salvage of extracellular GSH, giving rise
within the cell to oxidative processes such as lipid per-
It is well established that several species originating
from oxidative processes can interact with critical thiols
of proteins causing noxious effects, such as the impair-
ment of Ca2?-ATPases  or cytoskeletal alterations
. On the other hand, strong evidence has accumu-
lated that the same oxidizing species can also play a
nontoxic role as stimulants in transduction of prolifera-
tive signals, an effect likely due to their ability to interact
with oxidizable regions of growth factor receptors, pro-
tein kinases, and transcription factors [27–29]. Previous
studies showed that different pools of protein thiols can
be differentially affected by oxidants, depending on the
site where these originated . In this respect, the
peculiar location of GGT on cell plasma membrane
raises the possibility that oxidants arising from its activ-
ity may primarily affect the thiol redox status of proteins
located on the cell surface, possibly playing a physiolog-
ical role in the modulation of their functions.
Our study was designed to investigate the nature of
oxidants produced during GGT ectoactivity, and the ef-
fects possibly induced by modulation of GGT on the
thiol redox status of cell surface proteins. Results indi-
cate that an as yet unrecognized function of GGT might
indeed be related to the modulation of protein thiols at
the cell surface level, through the generation of hydrogen
peroxide and reactive thiols capable of effecting S-thio-
lation of cellular proteins.
MATERIALS AND METHODS
RPMI 1640 culture medium, N-ethylmaleimide
(NEM), N-hydroxymaleimide, iodoacetamide, diamide,
reduced glutathione, bovine kidney ?-glutamyl transpep-
tidase (EC 126.96.36.199), ?-glutamyl p-nitroanilide (GPNA),
tic acid), DL-dithiothreitol (DTT), glycyl-glycine, cys-
teinyl-glycine, L-serine, scopoletin, and type II horserad-
ish peroxidase (EC 188.8.131.52) were from Sigma (St.
Louis, MO, USA). N-(Biotinoyl)-N?-(iodoacetyl)ethyl-
enediamine (BIA) was from Molecular Probes (Eugene,
OR, USA). [35S]GSH (? 30 Ci/mmol) was purchased
from NEN Life Sciences (Boston, MA, USA). Trolox C
was from Fluka (Milan, Italy). Thymol-free liver catalase
(EC 184.108.40.206) was from Boehringer Mannheim (Monza,
Milan, Italy). Butylated hydroxytoluene (BHT) was from
Carlo Erba (Milan, Italy). Macromolecular hydroxy-
ethyl-deferoxamine (HES-DFO, consisting of 6 DFO
molecules per 50,000 D starch molecules) was donated
by Dr. Z. I. Cabantchik (The Hebrew University, Israel).
All other reagents used were of the highest available
Cell culture and incubation
Human histiocytic lymphoma U937 cells, expressing
significant levels of GGT ectoactivity (?15 mU/mg pro-
tein), were cultured in RPMI 1640 medium, containing
10% heat-inactivated FCS (Sigma), 2 mM L-glutamine
(Sigma), 0.25 ?g/ml amphotericin B and 50 ?g/ml gen-
tamycin, at 37°C, under 95% air:5% CO2, in a humidi-
fied atmosphere. Cells were maintained in logarithmic
growth by allowing each subculture to attain a density of
approximately 1.5–2.0 ? 106cells/ml before reseeding at
2 ? 105cells/ml.
For stimulation of cellular GGT activity, cell suspen-
sions (1 ? 106cells/ml) were incubated in the presence
of the substrate, GSH (100–200 ?M), and 1–2 mM
glycyl-glycine as the acceptor of ?-glutamyl moieties for
the transpeptidation reaction . Inhibition of GGT was
obtained by exposing cells to the specific, noncompeti-
tive inhibitor acivicin (130 ?M), for 4 h before experi-
ments, or alternatively, by performing experiments in the
presence of the specific competitive inhibitor serine/
624S. DOMINICI et al.
boric acid complex (10/20 mM) . In the conditions
employed (pretreatment of cells with 130 ?M for 4 h),
acivicin induced the 91.2 ? 1.5% inhibition of cell-
bound GGT activity (mean ? SEM of four experiments).
Where indicated, incubations also included catalase (100
?g/ml), Trolox C (1–3 mM), BHT (200 ?M), HES-DFO
(10–100 ?M), cysteinyl-glycine (10 ?M), and DTT
Fluorimetric determination of hydrogen
Extracellular H2O2production was determined as de-
scribed by Root et al. , by monitoring the decrease in
fluorescence of scopoletin during its oxidation catalyzed
by horseradish peroxidase (HRP). Cells were harvested
and resuspended at 3 ? 106/ml in HBSS without Ca2?
and Mg2?in the presence of 2 ?M scopoletin and 50 nM
HRP. Cells were then incubated at 37°C under continu-
ous gentle magnetic stirring, in the cuvette of a Perkin-
Elmer 650-10S fluorimeter set as follows: excitation/
emission wavelengths: 350/460 nm; slits: 4 nm;
sensitivity: 0.3 (arbitrary units). Incubations were per-
formed in the presence of GGT substrates (GSH and
gly-gly, for stimulated GGT conditions), and/or in the
presence of the specific inhibitors acivicin or serine/boric
acid complex (inhibited GGT conditions), as well as in
the presence of various antioxidant agents. Additional
incubations were performed in cell-free HBSS after the
addition of purified GGT or the GSH metabolite, cystei-
nyl-glycine. Calibration curves were carried out with
additions of standard H2O2to cell-free incubation me-
dium, in the range 0.1–5 ?M. Determinations were per-
formed adhering to the precautions suggested by
Determination of cell surface protein thiols by
SH groups in cell surface proteins were revealed by a
combined SH-biotinylation/ immunoblot/ECL proce-
dure, based on previously reported procedures . Fol-
lowing incubations, cell viability was routinely checked
by the Trypan blue exclusion test, and in all cases was
found to be higher than 95%. Cells were then washed
once, resuspended (1 ? 106/ml) in RPMI 1640 medium
without FCS, and exposed for 15 min at 37°C to the
noncell permeant thiol-labeleled BIA (final concentra-
tion of 100 ?M) added as small aliquots (10 ?l/ml) of a
10-mM stock solution of BIA in DMSO. Cells were then
washed in fresh medium, pelleted, and lysed by incubat-
ing for 45 min at 4°C with a small volume (approx. 10
?l/106cells) of a lysis buffer composed as follows: 0.5%
(v/v) Triton X-100, 5 mM Tris, 20 mM EDTA, 50 mM
NEM (pH 8.0). Cell lysates were then centrifuged for 15
min at 17,000 g at 4°C, and aliquots of the supernatants
were used for protein content determination. In some
experiments, the labeling of protein thiols was carried
out in cells that had been exposed to ultrasound in a
water bath for 5 min, or in cell lysates directly.
Additional lysate aliquots, containing equivalent
amounts of protein, were employed for separation of
cellular proteins by 10% SDS-PAGE, using biotinylated
high-molecular-mass standards in the range of 66.2–
116.0 kDa as molecular weight markers. Such separating
conditions were selected to obtain preliminary informa-
tion on the effects of GGT activity on the redox status of
cell surface proteins over a relatively wide mass range.
Gels were subsequently blotted onto nitrocellulose filter
paper (Sartorius, Go ¨ttingen, FRG) soaked in transfer
buffer (48 mM Tris, 39 mM glycine, 0.04% SDS [w/v],
and 20% methanol [v/v]). Efficiency of protein transfer
was routinely monitored by reversibly staining nitrocel-
lulose filters with Fast Green FCF (Sigma), dissolved (1
mg/ml) in 1% (v/v) glacial acetic acid. The labeling of
protein bands with BIA was revealed by exposing blots
to streptavidin-POD conjugate (Boehringer Mannheim,
FRG), followed by development with the BM Chemilu-
minescence Western Blotting kit (Boehringer Mann-
heim), using Kodak X-Omat AR films (exposure times
10 s to 5 min). Densitometric analysis of developed films
was carried out with a BioRad GS-690 imaging densi-
tometer apparatus, using the Molecular Analyst software.
Total extinction (Etot) values were obtained by semiau-
tomatic integration of all protein thiol bands in individual
lanes. Longitudinal densitometric profiles of individual
lanes were determined by manually selecting suitable,
flaw-free axes for scanning.
Cytochemical visualization of cell surface protein
thiols by laser-scanning confocal fluorescence
Procedures employed were essentially as previously
described . Briefly, U937 cell suspensions were re-
acted with 3-(N-maleimidylpropionyl) biocytin (MPB;
Molecular probes, Eugene, OR, USA), 25 ?g/ml in phos-
phate-buffered saline, for 30 min at room temperature.
Thoroughly washed cells were then smeared on clean
glass slides, postfixed with cold anhydrous acetone, and
treated with fluoresceinated extravidin (eaFITC). Confo-
cal fluorescence microscopy was performed with a Bio-
Rad MRC-500 confocal imaging system, equipped with
a 20-mW argon ion laser, using a Nikon Plan-Apo 60
(NA 1.4) oil immersion objective. Optical sections of
approximately 1.0-?m thickness were obtained at the
largest apparent cell diameter. Confocal images were
?-Glutamyl transpeptidase and cell surface protein thiols
obtained, stored, and processed employing the BioRad
COMOS software. All other machine settings were as
previously described .
Protein S-thiolation experiments
For the determination of binding to cellular proteins of
thiol compounds originating during GGT-mediated metab-
olism of extracellular GSH, U937 cells were incubated for
1 h in conditions of GGT stimulation or inhibition as
described above, in HBSS (3 ? 106cells/ml), in the pres-
ence of 200 ?M [35S]GSH (2 ?Ci/ml) and 2 mM glycyl-
glycine. At the end of the incubations, selected samples
were treated prior to cell lysis with 3 mM DTT for 15 min,
to eliminate S-thiolation of the cell surface proteins. All cell
samples were then washed twice with PBS and lysed as
previously described. Proteins of cell lysates were precipi-
(TCA); protein precipitates were filtered under vacuum
onto glass microfiber filters (Whatman GF/C, Maidstone,
UK), washed extensively with 7.5% TCA and cold ethanol,
air-dried, and counted in a scintillation ?-counter (Canberra
?-Glutamly transpeptidase activity was assayed with
the method of Sze et al. , involving the spectropho-
tometric measurement of p-nitroaniline released from the
substrate GPNA; this method is specific for GGT that
associates with the outer side of the plasma membrane
and metabolizes extracellular GSH. One unit of GGT
activity was defined as the amount of enzyme capable to
release 1 ?mol p-nitroaniline/min from the substrate
GPNA (Huseby and Stromme). Protein concentration
was determined with the method of Lowry, using bovine
serum albumin as a standard.
To characterize prooxidant reactions induced by cell-
bound GGT activity, U937 cells (possessing (?15 mU
GGT/mg protein at their surface, determined as reported
in Experimental Procedures) were incubated with the
GGT substrate GSH and cosubstrate ?-glutamyl acceptor
glycyl-glycine, and the production of hydrogen peroxide
was monitored fluorimetrically (HRP-mediated decrease
of scopoletin fluorescence) . Concentrations of sub-
strates were chosen in to approach the vmaxof GGT
enzyme activity , thus facilitating detection of hy-
drogen peroxide by the assay procedure. As reported in
Fig. 1A, the metabolism of GSH by GGT associated with
plasma membrane of U937 cells resulted in a sustained
production of hydrogen peroxide. The reaction occurred
in the extracellular environment, as shown by the fact
that catalase—which cannot penetrate the cell mem-
brane—would suppress it. Generation of H2O2did not
take place with cells in which GGT had been irreversibly
inhibited by the noncompetitive GGT inhibitor acivicin,
nor in the presence of the competitive GGT inhibitor,
serine-borate complex (Fig. 1B). Figure 1B also shows
that H2O2production was markedly decreased in the
presence of the vitamin E analogue Trolox C. H2O2
production was suppressed by addition of low concen-
trations of the noncell permeant iron chelator HES-DFO
(Fig. 1A), as well as by addition of a structurally-unre-
lated metal chelator, EDTA (50 ?M, not shown), thus
confirming the involvement of extracellular iron ions in
the reaction. As expected, H2O2production could also be
directly started by adding to the assay mixture cysteinyl-
glycine, the metabolite resulting from GGT-mediated
cleavage of GSH, although with a relatively low effi-
ciency as judged on a molar basis (Fig. 1C).
The possibility that part of the H2O2produced might
be simultaneously removed by cells, thus leading to an
underestimation of the process, was investigated by re-
placing cell-bound GGT activity with an equivalent
amount of enzyme units, in the form of purified GGT
protein. Indeed, as demonstrated in Fig. 1D, with the
purified enzyme, H2O2production was higher than with
the same amount of cell-bound GGT activity, indicating
the sequestration and/or metabolism of H2O2within the
Having documented that GGT-mediated metabolism
of GSH at the surface of U937 cells can generate hydro-
gen peroxide, the possibility that the latter could interfere
with the redox status of thiol groups contained in pro-
teins of the cell surface was then investigated. To this
aim, we used a procedure for the selective labeling of cell
surface protein thiols, based on the the thiol-specific
(BIA), as previously reported by our laboratory .
Figure 2 shows the results obtained with BIA-labeling of
control U937 cells. Total cellular protein thiols were
determined by carrying out the labeling incubation after
cell lysis. On the other hand, when BIA-labeling was
performed on intact cells, detectable protein thiols cor-
responded to only a minor fraction (approximately 25%)
of the total. This indicates that in intact cells, the label
indeed cannot react with proteins across the plasma
membrane, and that a preferential labeling of protein
thiols of the cell surface occurs. Accordingly, exposure
of cells to BIA-labeling after partial disruption of the
membrane barrier by short sonication resulted in the
increase of signal to an intermediate degree between
intact and lysed cells (Figs. 2B and 2C).
The procedure for determination of cell surface pro-
626S. DOMINICI et al.
tein thiols was further validated by exposing cells to
different thiol-specific reagents prior to BIA labeling. As
reported in Figure 3, BIA labeling allowed us to docu-
ment a loss of cell surface protein thiols with all reagents
tested. The latter exhibited different efficiencies (alleg-
edly reflecting their different degrees of polarity and/or
lipid/water partitioning). In particular, the loss of protein
thiols observed after treatment of cells with N-hydroxy-
maleimide—which is strictly non-cell-permeable, and
could thus only interact with proteins of the outer aspect
of cell membrane—demonstrated the ability of the la-
beling procedure to detect alterations occurring in pro-
teins of the cell surface. Interestingly, thiols were not
uniformly affected in all protein bands, because proteins
Fig. 1. GSH- and GGT-dependent extracellular production of hydrogen peroxide from U937 cells. Decrease of scopoletin fluorescence
in the presence of horseradish peroxidase. Individual traces have been offset to make it easier to compare their shapes and slopes.
Vertical bars indicate the fluorescence decrease corresponding to a concentration of 0.5 ?M H2O2, as established in preliminary
calibration experiments (see Experimental Procedures for details). (A) Production of H2O2by U937 cells (3 ? 106/ml) upon addition
of the substrate GSH (100 ?M) and the ?-glutamyl acceptor glycyl-glycine (gly-gly, 1 mM). H2O2production was dependent on the
extracellular availability of iron ions, as shown by the inhibition offered by adding the noncell permeable iron chelator HES-DFO (50
?M) or the structurally-unrelated metal chelator EDTA (50 ?M, not shown). Where indicated, catalase was added directly to the
incubation mixture (100 ?g/ml, final concentration). (B) Effects of GGT inhibitors and Trolox C on GSH-dependent H2O2production
by U937 cells. H2O2production was suppressed when cells were pretreated with the noncompetitive GGT inhibitor acivicin (130 ?M,
4 h) or monitored in the presence of the competitive GGT-inhibitor serine/boric acid complex (10/20 mM). H2O2production was also
prevented in the presence of the free radical scavenger Trolox C (1 mM). (C) Generation of H2O2induced by direct addition to the
incubation mixture of cysteinyl-glycine (10 ?M, final concentration), the product of GGT-mediated GSH metabolism. (D) H2O2
production rate by membrane-bound GGT activity of U937 cells (3 ? 106cells/ml), as compared to that observed in the presence of
an equivalent amount of purified GGT protein (corresponding to 8 mU/ml enzyme activity). Figure also shows that the addition to cells
of the GGT inhibitor serine/boric acid complex immediately results in the suppression of H2O2production. Typical experiments of 4
or 7 are shown; repeated experiments gave fully comparable results.
?-Glutamyl transpeptidase and cell surface protein thiols
Fig. 2. Determination of protein thiols after labeling of U937 cells with
the noncell permeant thiol label N-(biotinoyl)-N?-(iodoacetyl) ethyl-
enediamine (BIA) (see Experimental Procedures for details) (A) West-
ern blot. Lane 1: labeling of intact U937 cells; lane 2: labeling of lysed
cells; lane 3: labeling of cells exposed to ultrasound in a water bath (5
min). (B) Total extinction (Etot) values obtained by semi-automatic
integration of all protein thiol bands in individual lanes. (C) Densito-
metric profiles, obtained by longitudinal scanning of individual lanes.
One experiment of three is shown; repeated experiments gave fully
Fig. 3. BIA-labeling of cell surface protein thiols: effects of various
thiol-specific treatments (see Experimental Procedures for details). (A)
Western blot. Lane 1: untreated control; lane 2: 100 ?M NOHM; lane
3: 500 ?M iodoacetamide; lane 4: 100 ?M NEM; lane 5: 500 ?M
NEM; lane 6: 500 ?M diamide. All incubations were carried out at
37°C for 10 min. (B) Total extinction (Etot) values obtained by semi-
automatic integration of all protein thiol bands in individual lanes, as
from panel A. (C) Comparison of selected longitudinal densitometric
profiles (see legend to Fig. 3). One typical experiment of four is shown;
repeated experiments gave fully comparable results.
628 S. DOMINICI et al.
in selected molecular weight ranges proved less accessi-
ble to the reagents tested (Fig. 3C).
Thus, the procedure validated as above was employed
for the investigation of changes induced in cell surface
protein thiols by modulation of membrane-bound GGT
activity (Fig. 4 and Table 1). To stimulate GGT activity,
U937 cell suspensions were incubated with the substrate
GSH, in the absence or presence of the cosubstrate
?-glutamyl acceptor glycine-glycine. Stimulation of cell-
bound GGT activity by the addition of the substrate GSH
alone resulted in a significant decrease of protein thiols
(Fig. 4B and Table 1). Loss of protein thiols was en-
hanced in the presence of both GSH and the ?-glutamyl
acceptor gly-gly, for example, in conditions approaching
the vmaxof enzyme activity . Consistent with a role of
GGT-mediated reactions in modulation of cell surface
protein thiols, when the same substrates were added to
cells in which GGT had been more than 90% inhibited,
pretreatment with the specific irreversible inhibitor acivi-
cin, protein thiol loss was also significantly inhibited
(Fig. 4 and Table 1). Protein thiol loss was unaffected by
pretreatment of cells with the lipid antioxidant BHT,
ruling out a possible role of lipid peroxidation in the
process. On the other hand, a partial protection against
GGT-mediated protein thiol loss was offered by the free
radical scavenger Trolox C (Fig. 4B), confirming the
involvement in the process of radical reactions (Fig. 1B).
The modulation of GGT activity by the addition of
exogenous substrates GSH and glycyl-glycine was thus
reflected in alterations of the cell surface protein thiol
redox status. To verify whether GGT-dependent modu-
lation of cell surface protein thiols is a spontaneous
process, continuously operative in GGT-rich cells, fur-
ther experiments were performed in the absence of ex-
ogenously-added GSH or glycyl-glycine, and the effects
of inhibition of “basal” GGT activity were investigated.
As reported in Fig. 5, the irreversible inhibition of basal
GGT activity of cells with acivicin was sufficient to
produce an increase of detectable protein thiols at the cell
surface, as documented by both the BIA-labeling proce-
dure (Figs. 5A, 5B) and by confocal fluorescence mi-
croscopy (Figs. 5C, 5D). In particular, the latter proce-
dure allowed us to visualize a fluorescent rim at the
surface of cells exposed to acivicin, a pattern confirming
that GGT inhibition in fact induced an increase of MPB-
detectable protein thiols at the cell surface level.
To verify the possibility that the observed changes in
cell surface protein thiols may be caused by GGT-de-
rived H2O2, in additional experiments, GGT stimulation
was performed in the presence of catalase. As reported in
Table 1, catalase largely prevented GGT-mediated oxi-
dation of protein thiols, indicating the involvement of
H2O2in the process. Consistently, direct addition to cell
suspensions of micromolar concentrations of H2O2(i.e.,
Fig. 4. Modulation of cell surface protein thiols following modulation
of GGT activity. BIA labeling of intact U937 cells. (A) Western blot.
Lane 1: GGT full stimulation in the presence of Trolox C (3 mM); lane
2: cells incubated with GSH alone (200 ?M); lane 3: GGT full
stimulation; lane 4: untreated control cells; lane 5: GGT-inhibited cells,
incubated with GSH and gly-gly; lane 6: GGT full stimulation in the
presence of BHT (200 ?M). All incubations were carried out for 60
min at 37°C. “GGT full stimulation” was effected by incubating cells
with GSH (200 ?M) and gly-gly (2 mM). “GGT-inhibited cells” had
been treated with 130 ?M acivicin for 4 h prior to the experiment. (B)
Total extinction values (see legend to Fig. 4), as from panel A. (C)
Comparison of selected longitudinal densitometric profiles (see legend
to Fig. 3). One typical experiment of six is shown; repeated experi-
ments gave fully comparable results.
?-Glutamyl transpeptidase and cell surface protein thiols
in the range of those observed following GGT stimula-
tion, Fig. 1) indeed resulted in substantial loss of protein
thiols (Table 1).
The stimulation of GGT activity associated with
plasma membrane of U937 cells was accompanied by
production of hydrogen peroxide, and was able at the
same time to induce loss of thiol in proteins of the cell
surface. Protein thiol oxidation can occur by the forma-
tion of protein disulfides, involving protein cysteine res-
idues. On the other hand, protein thiol oxidation can also
result from the formation of “mixed disulfides,” involv-
ing protein cysteines and low molecular weight thiols
(e.g., GSH, cysteine, cysteamine), a process often termed
“protein S-thiolation” . The possibility that GGT-
mediated oxidation of protein thiols in U937 cells might
imply (at least in part) protein S-thiolation by reactive
low molecular weight thiols (such as cysteinyl-glycine
and cysteine originating from extracellular GGT-medi-
ated GSH metabolism) was thus investigated. To this
aim, experiments were performed in which [35S]GSH
was employed as a substrate for cell-bound GGT, and the
amounts of radioactivity associated with the protein frac-
tion were determined. The results are shown in Figure 6.
Stimulation of GGT activity by the addition of [35S]GSH
to cells resulted in the binding of [35S] to cellular protein,
and such [35S] binding was significantly increased when
cells were incubated with both GSH and glycyl-glycine,
such as in conditions of full GGT stimulation. On the
other hand, as expected, [35S] binding was largely pre-
vented in cells which had been pretreated with the irre-
versible GGT inhibitor acivicin. [35S] binding was also
decreased in the presence of the free radical scavenger,
Trolox C (Fig. 6). Finally, in GGT-stimulated cells,
treatment with DTT after incubation resulted in partial
reversal of the observed increase in [35S] binding, indi-
cating that GGT stimulation had indeed promoted the
formation of DTT-reductable disulfide bonds between
protein thiols and radiolabeled thiols deriving from
GGT-mediated metabolism of [35S]GSH (Fig. 6).
The main function of the GGT-mediated metabolism
of extracellular GSH appears to lie in the recovery of
cysteine, whose adequate supply is critical for protein
synthesis, especially in rapidly dividing neoplastic cells
. On the other hand, a further effect of GGT-medi-
ated GSH metabolism appears to lie in the induction of
oxidizing processes [21–24]. The latter can be explained
by the fact that GGT—by effecting the cleavage of the
?-glutamyl moiety from GSH molecule—generates cys-
teinyl-glycine (a thiol provided with much higher reac-
tivity as compared to GSH). Due to a pKa lower than that
of GSH , at near-neutral pH, cysteinyl-glycine will
be mostly in its dissociated form gly-cys-S?(thiolate
anion). In principle, gly-cys-S?thiolate anions would be
capable to start redox interactions, for example, with iron
ions present in the medium, leading to the formation of
reactive oxygen species and thiyl radicals, according to
the following sequence:
?-glutamate ? gly-cys-S?? H?
¡ gly-cys-S•? Fe2?
¡ Fe3?? O2
•?? H2O O
¡ O2? H2O2
Table 1. Gamma-Glutamyl Transpeptidase-Dependent Modulation of Thiols in Surface
Proteins of U937 Cells
Cell surface protein thiols
?GSH 200 ?M
?GSH 200 ?M ? gly-gly 2 mM (?GGT full stimulation)
?GSH ? gly-gly (cells pretreated with acivicin 130 ?M)
?GSH ? gly-gly ? catalase (100 ?g/ml)
58.0 ? 13.2a
44.7 ? 9.05a
98.1 ? 17.4b
86.5 ? 0.6b
74.1 ? 7.0
Percent values were calculated from the total extinction (Etot) values obtained in individual
experiments by semiautomatic integration of BIA-labeled protein bands (see Materials and
Methods for details of incubations and densitometric analysis).
Data reported are means ? SEM of three to five experiments.
Statistical significance of data (p ? .05) was assessed by analysis of variance with Student-
Newman-Keuls test for multiple comparisons.
aSignificantly different from control value.
bSignificantly different from GGT full-stimulation value.
630S. DOMINICI et al.
Interactions of gly-cys-S?with metal ions would be
favored also due to the loss of ?-glutamate, whose pres-
ence in GSH was shown to prevent the reduction of
transition metal ions by the cysteine -SH group [38,39].
Data shown in Fig. 1 indeed support the sequence of
reactions envisaged above. In fact, the exposure of U937
cells to conditions of stimulated GGT activity resulted in
the production of detectable amounts of H2O2. The pro-
cess was occurring in the extracellular space, as indicated
by the fact that catalase, which is non-cell permeable,
would suppress it (Fig. 1A), and involved the generation
of free radicals, as suggested by the suppressing effect of
Trolox C (Fig. 1B). The scavenging ability of Trolox C
has been documented with several free radical species,
including superoxide . The central role played by
GGT activity in the process is demonstrated by the fact
that no H2O2production was observed with cells in
which GGT activity had been inhibited by pretreatment
with the noncompetitive, irreversible GGT inhibitor
acivicin (Fig. 1B). The finding was further confirmed
with cells incubated in the presence of the competitive
GGT inhibitor serine-borate complex (Fig. 1B), whose
mechanism is distinct from that of acivicin . Further-
more, the involvement of extracellular iron ions in the
process is indicated by the suppressive effect of the
noncell permeable iron chelator HES-DFO (Fig. 1A), as
well as of EDTA. No exogenous iron was actually in-
cluded in the incubation mixture; however, it is well
established that standard buffer solutions and cell culture
media do contain trace amounts of contaminant iron, in
the low micromolar range .
Thus, GGT activity associated with plasma membrane
of U937 cells appears to act as a potential source of H2O2
in the extracellular space. Being freely membrane-per-
meable, H2O2can diffuse into the cells themselves. This
is documented by data reported in Fig. 1D. In fact, H2O2
production detectable with cell-bound GGT was lower
than that detected with the same amount of purified
enzyme activity, indicating that part of the H2O2formed
by cells actually diffuses into the cells themselves, where
it becomes undetectable by the scopoletin-HRP assay
system , and can be metabolized by cellular GSH-
Fig. 5. Increase of cell surface protein thiols following irreversible
inhibition of GGT activity by acivicin. (A) BIA labeling of intact U937
cells and Western blot. Lane 1: control cells; lane 2: GGT-inhibited
cells (24-h exposure to 130 ?M acivicin prior to the experiment). One
typical experiment of three is shown; repeated experiments gave fully
comparable results. (B) Longitudinal densitometric profiles of samples
reported in panel A. (C, D) False color (see scale bar in panel C)
confocal imaging of fluorescence levels detected in U937 cells labeled
for cell surface protein thiols with MPB/eaFITC (see Experimental
Procedures for details). Imaging parameters were adjusted to minimize
visualization of control cells (panel C), thus making the increase of
fluorescence in GGT-inhibited cells immediately apparent (panel D).
GGT inhibition was obtained by exposing U937 cells to 130 ?M
acivicin for 24 h prior to the experiment. One typical experiment of
four is shown; repeated experiments gave fully comparable results. Bar
in panel C corresponds to 10 ?m.
Fig. 6. S-thiolation of cellular protein following modulation of plasma
membrane-bound GGT activity. U937 cells were incubated for 1 h in
HBSS in the presence of [35S]GSH alone, 200 ?M, or with both
[35S]GSH and the GGT cosubstrate gly-gly, 2 mM (? GGT full
stimulation). Where indicated, cells had been pretreated with the GGT
inhibitor acivicin (130 ?M, 4 h) prior to incubation with GSH and
gly-gly. Further samples were incubated in the presence of [35S]GSH,
gly-gly, and Trolox C (3 mM). Following incubations, additional
[35S]GSH/gly-gly-treated samples (intact cells) were exposed to 3 mM
DTT for 15 min to reduce disulfide bonds, thus eliminating cell surface
S-thiolation. Data shown are means ? SEM of three separate experi-
ments, and are expressed as percent of the GGT full stimulation value.
(*) p ? .05, as compared to GGT stimulation, and as assessed by
analysis of variance with Student-Newman-Keuls test for multiple
?-Glutamyl transpeptidase and cell surface protein thiols
peroxidase and catalase. Also, the lower amounts of
H2O2detectable in the presence of cells could suggest
that part of H2O2produced is promptly reacting with
molecular targets, such as at the cell surface.
It is well established that oxidants can interact with
critical thiols of proteins. While such interactions are
capable of producing toxic effects to the cell [25,26],
recent studies have provided compelling evidence that
the same oxidizing species (at lower concentrations) can
also play a nontoxic role by modulating the function of
growth factor receptors, protein kinases, and transcrip-
tion factors, through alteration of oxidizable regions in
their structure [27–29]. Conceivably, a primary target for
the action of prooxidants generated extracellularly dur-
ing GGT activity would be given by thiols of proteins
located at the cell surface. Therefore, we developed and
validated a procedure for the selective labeling of thiols
of cell surface proteins (Figs. 2 and 3). Indeed, the results
obtained (Fig. 4, Table 1) revealed the occurrence of a
GGT-dependent oxidation of thiol groups in surface pro-
teins of U937 cells. Protein thiol oxidation was in fact
increased following stimulation of GGT activity, while
the process was prevented after its inhibition. The in-
volvement of hydrogen peroxide in the process is indi-
cated by the fact that protein thiol oxidation was signif-
icantly prevented by catalase, and that a protein thiol
loss, although not significant, was produced by micro-
molar concentrations of H2O2(Table 1).
It should be emphasized that, although the suscepti-
bility of protein thiols to the oxidizing action of hydro-
gen peroxide is a known and well-assessed phenomenon,
a novel notion supported by data reported in the present
study is given by the identification of plasma membrane
GGT activity as an unrecognized cellular source of H2O2
involved in modulation of protein thiol redox status. In
fact, as shown in Fig. 5, GGT-dependent oxidation of
cell surface proteins appears to be a continuous, basal
process in U937 cells, because the inhibition of basal,
unstimulated GGT activity resulted in an increase of
detectable protein thiols at the cell surface. It can be
envisaged that GGT inhibition, by removing a basal
GGT-dependent prooxidant action, produces an actual
shift in the cell redox equilibrium, allowing protein thiols
of the cell surface to regain a more reduced status. In
GGT-rich cells, surface proteins appear thus to be con-
tinuously exposed to a GGT-dependent oxidant stress,
which maintains their thiols partially oxidized. In this
perspective, GGT activity of U937 cells could play a
physiological role, as one of the determinants of the
redox status of cell surface protein thiols.
Different mechanisms can be envisaged to concur in
determining GGT-mediated protein thiol oxidation. On
one hand, direct interaction of hydrogen peroxide with
protein thiols can take place, a process mainly resulting
in the oxidation of thiol groups to sulfenic and sulfinic
acid residues [43,44]. On the other hand, several studies
have documented the ability of H2O2to stimulate the
process of “S-thiolation” [45–48], for example, the for-
mation of “mixed” disulfide bonds involving cysteine
residues on protein and low molecular weight thiols.
Indeed, data reported in Fig. 6 document that stimulation
of GGT activity does result in the promotion of S-
thiolation of cellular proteins, such as in the formation of
(DTT-reversible) mixed disulfide bonds between protein
thiols and metabolites of extracellular GSH. As intact
cells were exposed to DTT action, the reversal of [35S]
binding observed, which was not complete, can be inter-
preted as indicating the extent of [35S] binding to pro-
teins of the cell surface. On the other hand, the DTT-
resistant labeling can be interpreted as representing
S-thiolation of intracellular proteins, and/or the internal-
ization of S-thiolated surface proteins to cellular com-
partments not rapidly accessible by DTT in intact cells.
S-thiolation of intracellular proteins could be promoted
in our conditions by intracellular diffusion of GGT-
derived hydrogen peroxide. Also, the possibility cannot
be excluded that non-DTT-reversible [35S] binding ob-
served in our experiments may represent a putative in-
corporation of [35S] into cellular proteins. However, dif-
fering from other published studies, inhibition of protein
synthesis could not be carried out in our S-thiolation
experiments, because it would have affected GGT activ-
ity itself, due to the unusually rapid turnover of GGT
protein—up to 50% in 90 min .
In addition to stimulation by H2O2, an alternative
mechanism for the observed protein S-thiolation could
be mediated through the so-called “GSH-oxidase” activ-
ity of GGT. Variable amounts of oxidized glutathione
(GSSG) are in fact produced during GGT enzymatic
activity , and formation of protein mixed disulfides
mediated by GSSG (“protein-S-glutathiolation”) has
been described in several experimental models .
However, this aspect was not investigated in the present
It has been envisaged that S-thiolation and de-thiola-
tion processes can have regulatory effects on function of
a wide number of proteins and enzymes, as cysteine
residues are located in the active sites of several essential
enzymes . In a more specific perspective, the rele-
vance to tumor biology of prooxidant reactions and pro-
tein S-thiolation mediated by GGT appears to lie in the
possibility that these processes are continuously, basally
operative in cells expressing GGT activity at their sur-
face. The in vitro experimental conditions employed in
the present study included relatively high concentrations
of reagents (GSH, gly-gly). These were chosen based on
the described Kmvalues of GGT for its substrates, in
order to optimize conditions for GGT activity and thus
632S. DOMINICI et al.
enhance GGT-dependent prooxidant reactions beyond
the threshold of detectability with the available methods.
On the other hand, GGT-mediated prooxidant reactions
can be sustained in in vivo conditions as well. In fact,
with respect to the role of ?-glutamyl acceptor, which
was sustained by gly-gly in the present study, this can be
played in vivo by a number of different amino acids and
small peptides, including GSH itself and its metabolites
. Furthermore, concerning the availability of extracel-
lular GSH, it has been documented that a continuous
GSH efflux, through specific membrane transporters,
takes place from mammalian cells , and that the
extracellular GSH thus originated is indeed continuously
reabsorbed by cells after its cleavage by GGT . As it
plays a central role in such continuous “GSH cycling”
across the plasma membrane, it is therefore likely that—
although at levels conceivably lower than those observed
in the present study—GGT-dependent H2O2production
and protein S-thiolation can take place spontaneously in
in vivo conditions as well.
The biological significance of such an as yet unrec-
ognized physiological function of GGT in tumor cells
may be multifold. In the first place, GGT-mediated
prooxidant reactions might interfere with the cellular
responsivity to exogenous proliferative signals, because
it has been documented that oxidative alterations in crit-
ical cysteine residues in receptor proteins at the cell
surface can modify the ligand-binding affinity and the
activation status of important growth factor receptors,
such as EGFR [51,52] b-FGFR , and insulin recep-
tors . In addition, oxidizing species can play a role
within the cell by modulating the transduction of prolif-
erative signals, due to their ability to interact with oxi-
dizable regions of protein kinases and transcription fac-
tors [27–29,55], as well as of the p53 protein ; freely
cell-permeable, extracellular hydrogen peroxide pro-
duced during GGT activity can participate in these pro-
cesses. Indeed, in a previous study from our laboratory,
glutathione was shown to affect proliferation of ovarian
cancer cells through a GGT-dependent mechanism, in-
volving both production of hydrogen peroxide and oxi-
dation of protein thiols . Furthermore, a role of GGT
has been documented in the apoptotic death of lympho-
blastoid cells , and we have shown that GGT-depen-
dent extracellular production of H2O2is essential for the
maintainance of cell proliferation and prevention of ap-
optosis in U937 cells . A link could thus be estab-
lished between GGT expression, generation of oxidant
species, and functioning of cellular mechanisms for con-
trol of cell proliferation and apoptosis . Finally, it has
been hypothesized that conversion of protein sulfhydryls
to mixed disulfides (protein S-thiolation) might represent
a way of protecting them against oxidative damage .
As treatment of malignant tumors often involves use of
radiation or prooxidant chemotherapeutics whose effects
are largely dependant upon the generation of free radi-
cals, GGT-mediated S-thiolation reactions in tumor cells
might then represent a factor in resistance against se-
lected drugs and radiation therapy.
In conclusion, data reported in the present study pro-
vide evidence in favor of a crucial role of plasma mem-
brane-bound GGT enzymatic activity in the modulation
of the redox status of cellular protein thiols, with special
reference to proteins of the cell surface. This action
appears to be mediated through the production of hydro-
gen peroxide on one hand, and of reactive thiol metab-
olites of GSH on the other hand. Interactions of these
compounds with each other, and with cellular proteins,
lead to redox modulation of protein thiols, with potential
reflections on the proliferative attitude, and possibly on
the resistance to oxidant drugs, of tumor cells expressing
significant levels of plasma membrane GGT activity.
Acknowledgements — We are indebted to Dr. Z. I. Cabantchik (The
Hebrew University, Jerusalem, Israel) for the kind gift of HES-DFO.
This study was supported by the Associazione Italiana Ricerca sul
Cancro (A.I.R.C., Italy). Additional funds were derived from A.I.C.R.
(U.K.) and the Italian Ministry for University and Scientific Research
 Hagen, T. M.; Aw, T. Y.; Jones, D. P. Glutathione uptake and
protection against oxidative injury in isolated kidney cells. Kidney
Int. 34:74–81; 1988.
 Hagen, T. M.; Bai, C.; Jones, D. P. Stimulation of glutathione
absorption in rat small intestine by alpha-adrenergic agonists.
FASEB J. 5:2721–2727; 1991.
 Tate, S. S.; Meister, A. ?-Glutamyl transpeptidase from kidney.
Meth. Enzymol. 113:400–419; 1985.
 Lu, S. C.; Sun, W.-M.; Jian, Y.; Ookhtens, M.; Sze, G.; Kaplow-
itz, N. Role of two recently cloned rat liver GSH transporters in
the ubiquitous transport of GSH in mammalian cells. J. Clin.
Invest. 97:1488–1496; 1996.
 Forman, H. J.; Liu, R.-M.; Tian, L. Glutathione cycling in oxi-
dative stress. Lung Biol. Health Dis. 105:99–121; 1997.
 Hanigan, M. H.; Frierson, H. F. Jr.; Brown, J. E.; Lovell, M. A.;
Taylor, P. T. Human ovarian tumors express ?-glutamyl transpep-
tidase. Cancer Res. 54:286–290; 1994.
 Paolicchi, A.; Pompella, A.; Tonarelli, P.; Gadducci, A.; Genaz-
zani, A. R.; Zunino, F.; Pratesi, G.; Tongiani, R. Gamma-glutamyl
transpeptidase activity in human ovarian carcinoma. Anticancer
Res. 16:3053–3058; 1996.
 Munjal, D. D. Concurrent measurements of carcinoembryonic
antigen, glucose phosphate isomerase, ?-glutamyl transferase and
lactate dehydrogenase in malignant, normal adult and fetal colon
tissues. Clin. Chem. 26:1809–1812; 1980.
 Murata, J.; Ricciardi-Castagnoli, P.; Dessous L’Eglise Mange, P.;
Martin, F.; Juillerat-Jeanneret, L. Microglial cells induce cyto-
toxic effects toward colon carcinoma cells: measurement of tumor
cytotoxicity with a gamma-glutamyl transpeptidase assay. Int. J.
Cancer 70:169–174; 1997.
 Blair, S. L.; Heerdt, P.; Sachar, S.; Abolhoda, A.; Hochwald, S.;
Cheng, H.; Burt, M. Glutathione metabolism in patients with
non-small cell lung cancers. Cancer Res. 57:152–155; 1997.
 Tsutsumi, M.; Sakamuro, D.; Takada, A.; Zang, S. C.; Furukawa,
T.; Taniguchi, N. Detection of a unique gamma-glutamyl
transpeptidase messenger RNA species closely related to the
?-Glutamyl transpeptidase and cell surface protein thiols
development of hepatocellular carcinoma in humans: a new can-
didate for early diagnosis of hepatocellular carcinoma. Hepatol-
ogy 23:1093–1097; 1996.
 Hochwald, S. N.; Harrison, L. E.; Rose, D. M.; Anderson, M.;
Burt, M. E. Elevation of glutathione and related enzyme activities
in high-grade and metastatic extremity soft tissue sarcoma. Ann.
Surg. Oncol. 4:303–309; 1997.
 Supino, R.; Mapelli, E.; Sanfilippo, O.; Silvestro, L. Biological
and enzymatic features of human melanoma cell clones with
different invasive potential. Melanoma Res. 2:377–384; 1992.
 Ta ¨ger, M.; Ittenson, A.; Franke, A.; Frey, A.; Gassen, H. G.;
Ansorge, S. ?-Glutamyl transpeptidase - cellular expression in
populations of normal human mononuclear cells and patients
suffering from leukemias. Ann. Hematol. 70:237–242; 1995.
 Morell, A.; Losa, G.; Carrel, S.; Heumann, D.; von Fliedner, V. E.
Determination of ectoenzyme activities in leukemic cells and in
established hematopoietic cell lines. Am. J. Hematol. 21:289–
 Tew, K. D.; Monks, A.; Barone, L.; Rosser, D.; Akerman, G.;
Montali, J. A.; Wheatmley, J. B.; Schmidt, D. E. Jr. Glutathione-
associated enzymes in the human cell lines of the National Cancer
Institute Drug Screening Program. Mol. Pharmacol. 50:149–159;
 Warren, B. S.; Naylor, M. F.; Winberg, L. D.; Yoshimi, N.;
Volpe, J. P.; Gimenez-Conti, I.; Slaga, T. J. Induction and inhi-
bition of tumor progression. Proc. Soc. Exp. Biol. Med. 202:9–15;
 Li, Y.; Lieberman, M. W. Two genes associated with liver cancer
are regulated by different mechanisms in rasT24 transformed liver
epithelial cells. Oncogene 4:795–798; 1989.
 Braun, L.; Goyette, M.; Yaswen, P.; Thompson, N. L.; Fausto, N.
Growth in culture and tumorigenicity after transfection with the
ras oncogene of liver epithelial cells from carcinogen-treated rats.
Cancer Res. 47:4116–4124; 1987.
 Godwin, A. K.; Meister, A.; O’Dwyer, P. J.; Huang, C. S.;
Harnilton, T. C.; Anderson, M. E. High resistance to cisplatin in
human ovarian cell lines is associated with marked increase of
glutathione synthesis. Proc. Natl. Acad. Sci. USA 89:3070–3074;
 Stark, A.-A.; Pagano, D. A.; Arad, A.; Siskindovitch, S.; Zeiger,
E. Effect of pH on mutagenesis by thiols in Salmonella typhi-
murium TA102. Mutat. Res. 224:89–94; 1989.
 Stark, A.-A.; Zeiger, E.; Pagano, D. A. Glutathione metabolism
by ?-glutamyl transpeptidase leads to lipid peroxidation: charac-
terization of the system and relevance to hepatocarcinogenesis.
Carcinogenesis 14:183–189; 1993.
 Pompella, A.; Paolicchi, A.; Dominici, S.; Comporti, M.; Tongi-
ani, R. Selective colocalization of lipid peroxidation and protein
thiol loss in chemically induced hepatic preneoplastic lesions: the
role of ?-glutamyl transpeptidase activity. Histochem. Cell Biol.
 Paolicchi, A.; Tongiani, R.; Tonarelli, P.; Comporti, M.; Pom-
pella, A. Gamma-glutamyl transpeptidase-dependent lipid peroxi-
dation in isolated hepatocytes and HepG2 hepatoma cells. Free
Radic. Biol. Med. 22:853–860; 1997.
 Moore, R. B.; Bamberg, A. D.; Wilson, L. C.; Jenkins, L. D.;
Mankad, V. N. Ascorbate protects against tert-butyl hydroperox-
ide inhibition of erythrocyte membrane Ca2??Mg2?-ATPase.
Arch. Biochem. Biophys. 278:416–424; 1990.
 Mirabelli, F.; Salis, A.; Perotti, M.; Taddei, F.; Bellomo, G.;
Orrenius, S. Alterations of surface morphology caused by the
metabolism of menadione in mammalian cells are associated with
the oxidation of critical sulfhydryl groups in cytoskeletal proteins.
Biochem. Pharmacol. 37:3423–3427; 1988.
 Lander, H. M. An essential role for free radicals and derived
species in signal transduction. FASEB J. 11:118–124; 1997.
 Suzuki, Y. J.; Forman, H. J.; Sevanian, A. Oxidants as stimulators
of signal transduction. Free Radic. Biol. Med. 22:269–285; 1997.
 Monteiro, H. P.; Stern, A. Redox modulation of tyrosine phos-
phorylation-dependent signal transduction pathways. Free Radic.
Biol. Med. 21:323–333; 1996.
 Pompella, A.; Romani, A.; Benedetti, A.; Comporti, M. Loss of
membrane protein thiols and lipid peroxidation in allyl alcohol
hepatotoxicity. Biochem. Pharmacol. 41:1255–1259; 1991.
 Huseby, N. E.; Stro ¨mme J. H. Practical points regarding routine
determination of ?-glutamyl transferase (?-GT) in serum with a
kinetic method at 37°C. Scand. J. Clin. Lab. Invest. 34:357–361;
 Root, R. K.; Metcalf, J.; Oshino, N.; Chance, B. H2O2release
from human granulocytes during phagocytosis. J. Clin. Invest.
 Donahue, W. F. Interference in fluorometric hydrogen peroxide
determination using scopoletin-horseradish peroxidase. Environ.
Toxicol. Chem. 17:783–787; 1988.
 Pompella, A.; Cambiaggi, C.; Dominici, S.; Paolicchi, A.; Tongi-
ani, R.; Comporti, M. Single-cell investigation by laser scanning
confocal microscopy of cytochemical alterations resulting from
extracellular oxidant challenge. Histochem. Cell Biol. 105:173–
 Sze, G.; Kaplowitz, N.; Ookhtens, M.; Lu, S. C. Bidirectional
membrane transport of intact glutathione in HepG2 cells. Am. J.
Physiol. 28:G1128–G1134; 1993.
 Thomas, J. A.; Chai, Y.-C.; Jung, C.-H. Protein S-thiolation and
dethiolation. Meth. Enzymol. 233:385–395; 1994.
 Hanigan, M. H.; Ricketts, W. A. Extracellular glutathione is a
souorce of cysteine for cells that express gamma-glutamyl
transpeptidase. Biochemistry 32:6302–6306; 1993.
 Spear, N.; Aust, S. D. Thiol-mediated NTA-Fe(III) reduction and
lipid peroxidation. Arch. Biochem. Biophys. 312:198–202; 1994.
 Paolicchi, A.; Minotti, G.; Tonarelli, P.; Tongiani, R.; De Cesare,
D.; Mezzetti, A.; Dominici, S.; Comporti, M.; Pompella, A.
Gamma-glutamyl transpeptidase-dependent iron reduction and
LDL oxidation - a potential mechanism in atherosclerosis. J. In-
vest. Med. 47:151–160; 1999.
 Kundu, S. C.; Willson, R. L. Thiyl (sulfhydryl/thiol) free radical
reactions, vitamins, ?-carotene, and superoxide dismutase in ox-
idative stress: design and interpretation of enzymatic studies.
Meth. Enzymol. 251:69–86; 1995.
 Tate, S. S.; Meister, A. Serine-borate complex as a transition-state
inhibitor of ?-glutamyl transpeptidase. Proc. Natl. Acad. Sci. USA
 Halliwell, B.; Gutteridge, J. M. C. Free radicals in biology and
medicine. Oxford: Clarendon Press; 1989.
 Radi, R.; Beckman, J. S.; Bush, K. M.; Freeman, B. A. Peroxyni-
trite oxidation of sulfhydryls - The cytotoxic potential of super-
oxide and nitric oxide. J. Biol. Chem. 266:4244–4250; 1991.
 Quesada, A. R.; Byrnes, R. W.; Krezoski, S. O.; Petering, D. H.
Direct reaction of H2O2with sulfhydryl groups in HL-60 cells:
zinc-metallothionein and other sites. Arch. Biochem. Biophys.
 Park, E. M.; Thomas, J. A. S-thiolation of creatine kinase and
glycogen phosphorylase b initiated by partially reduced oxygen
species. Biochim. Biophys. Acta 964:151–160; 1988.
 Ravichandran, V.; Seres, T.; Moriguchi, T.; Thomas, J. A.;
Johnston, R. B. Jr. S-thiolation of glyceraldehyde-3-phosphate
dehydrogenase induced by the phagocytosis-associated respira-
tory burst in blood monocytes. J. Biol. Chem. 269:25010–25015;
 Moriguchi, T.; Seres, T.; Ravichandran, V.; Sasada, M.; Johnston,
R. B. Jr. Diamide primes neutrophils for enhanced release of
superoxide anion: relationship to S-thiolation of cellular proteins.
J. Leukocyte Biol. 60:191–198; 1996.
 Seres, T.; Ravichandran, V.; Moriguchi, T.; Rokutan, K.; Thomas,
J. A.; Johnston, R. B. Jr. Protein S-thiolation and dethiolation
during the respiratory burst in human monocytes - A reversible
post-translational modification with potential for buffering the
effects of oxidant stress. J. Immunol. 156:1973–1980; 1996.
 Mathis, G. A.; Wyss, P. A.; Schuetz, E. G.; Hughey, R. P.; Sirica,
A. E. Expression of multiple proteins structurally related to ?-glu-
tamyl transpeptidase in non-neoplastic adult rat hepatocytes in
vivo and in culture. J. Cell. Physiol. 146:234–241; 1991.
 Griffith, O. W.; Tate, S. S. The apparent glutathione-oxidase
634S. DOMINICI et al.
activity of ?-glutamyl transpeptidase. J. Biol. Chem. 255:5011–
 Clark, S.; Konstantopoulos, N. Sulfhydryl agents modulate insu-
lin- and epidermal growth factor (EGF)-receptor kinase via reac-
tion with intracellular receptor domains: differential effects on
basal versus activated receptors. Biochem. J. 292:217–223; 1993.
 Huang, R.P.; Wu, J. X.; Fan, Y.; Adamson, E. D. UV activates
growth factor receptors via reactive oxygen intermediates. J. Cell
Biol. 133:211–220; 1996.
 Herbert, J. M.; Bono, F.; Savi, P. The mitogenic effect of H2O2for
vascular smooth muscle cells is mediated by an increase of the
affinity of basic fibroblast growth factor for its receptor. FEBS
Lett. 395:43–47; 1996.
 Schmid, E.; El Benna, J.; Galter, D.; Klein, G.; Dro ¨ge, W. Redox
priming of the insulin receptor ?-chain associated with altered
tyrosine kinase activity and insulin responsiveness in the absence
of tyrosine autophosphorylation. FASEB J. 12:863–870; 1998.
 Flohe ´, L.; Brigelius-Flohe ´, R.; Saliou, C.; Traber, M. G.;Packer,
L. Redox regulation of NF-kappa B activation. Free Radic. Biol.
Med. 22:1115–1126; 1997.
 Sun, Y.; Oberley, L. W. Redox regulation of transcriptional
activators. Free Radic. Biol. Med. 21:335–348; 1996.
 Perego, P.; Paolicchi, A.; Tongiani, R.; Pompella, A.; Tonarelli,
P.; Carenini, N.; Romanelli, S.; Zunino, F. The cell-specific
anti-proliferative effect of reduced glutathione is mediated by
?-glutamyl transpeptidase-dependent extracellular pro-oxidant re-
actions. Int. J. Cancer 71:246–250; 1997.
 Graber, R.; Losa, G. A. Apoptosis in human lymphoblastoid cells
induced by acivicin, a specific ?-glutamyltransferase inhibitor.
Int. J. Cancer 62:443–448; 1995.
 Del Bello, B.; Paolicchi, A.; Comporti, M.; Pompella, A.; Mael-
laro, E. Hydrogen peroxide produced during gamma-glutamyl
transpeptidase activity is involved in prevention of apoptosis and
maintenance of cell proliferation in U937 cells. FASEB J. 13:69–
 Coan, C.; Ji, J.-Y.; Hideg, K.; Mehlhorn, R. J. Protein sulfhydryls
are protected from irreversible oxidation by conversion to mixed
disulfides. Arch. Biochem. Biophys. 295:369–378; 1992.
b-FGFR—basic fibroblast growth factor receptor
EGFR—epidermal growth factor receptor
FCS—fetal calf serum
HBSS—Hanks’ balanced salts solution
SDS—sodium dodecyl sulfate
?-Glutamyl transpeptidase and cell surface protein thiols