Content uploaded by Takashi Agui
Author content
All content in this area was uploaded by Takashi Agui
Content may be subject to copyright.
Available via license: CC BY 4.0
Content may be subject to copyright.
Vol. 265, No. 1. Issue of January 5, pp. 334-340,199O
Printed in U.S.A.
Ecto-ATPase Activity in Cytolytic T-lymphocytes
PROTECTION FROM THE CYTOLYTIC EFFECTS OF EXTRACELLULAR ATP*
(Received for publication, May 26,1989)
Antonio FilippiniS, Rolf E. Taffs, Takashi Agui, and Michail V. Sitkovsky
From the Laboratory of Immunology, National Institute of Allergy and Infectious Diseases, National Institutes of Health,
Bethesda, Maryland 26892
Addition of ATP or ATP analog to the incubation
media is shown to result in cell death in experiments
with different cultured cell lines as evidenced by the
results of several independent assays, both in the ab-
sence or presence of extracellular Ca’+. Cytolytic T-
lymphocyte (CTL) clone itself was not only resistant to
cytolytic effects of ATP, but was able to “rescue” an-
tigen-nonbearing SICr-labeled cells from lytic effects
of extracellular ATP (but not from lytic effects of
adenosine 5’-thiotriphosphate) when present during
assay. To test whether the resistance of CTL to ATP is
due to a high activity of ecto-ATPase, four independent
assays of ATPase activity were utilized to demonstrate
the presence and relatively high activity of the ecto-
ATPase(s) on CTL surface. Studies of substrate speci-
ficity of CTL ecto-ATPase suggest that there is more
than one nucleoside 5’-triphosphatase on the surface
of CTL. The enzyme(s) activity is Ca2+- and M8+-
dependent and in this respect is similar to recently
described hepatic cells ecto-ATPase. We tested effects
of known ATP-binding site-specific reagents fluores-
cein 5’-isothiocyanate (FITC) and B’-fluorosulfonyl-
benzoyladenosine (FSBA) to find covalent modification
procedures to be used in studies of functional role of
ecto-ATPase. FSBA, but not FITC, inhibits lymphocyte
ecto-ATPase but addition of ATP together with FSBA
protects e&o-ATPase activity. Inactivation of CTL
ecto-ATPase by pretreatment with FSBA makes CTL
susceptible to lytic effects of extracellular ATP, as was
hypothesized for the functional role of this enzyme in
CTL.
Various cells release adenosine 5’-triphosphate (ATP) into
the media in response to activating stimuli (l-6). Extracellu-
lar ATP profoundly effects cellular functions (3), as well as
cell membrane properties and intracellular biochemical reac-
tions (reviewed in Ref. 7). Lymphocytes and other cell types
(8-10) have on the surface specific binding sites for nucleo-
tides. ATP at 1 mM concentration can induce mitogenesis in
immature murine thymocytes (11, 12) and histamine release
from mast cells (13). Concentrations of ATP as low as 500
pM are able to affect intracellular Ca*+ in thymocytes (14).
ATP also interacts with P,-purinergic receptors and interferes
with early biochemical events of cell activation (15). The
tetrabasic anion ATP4- (present as a minor equilibrium com-
ponent in divalent cation containing solutions of ATP) can
* The costs of publication of this article were defrayed in part by
the payment of page charges. This article must therefore be hereby
marked
“aduertisement”
in accordance with 18 USC. Section 1734
solely to indicate this fact.
$ Supported by a grant from the Italian Association for Cancer
Research.
permeabilize many different cell types that have receptors for
ATP4- by forming pores and/or lesions in the plasma mem-
brane (15-21).
A prevalent model of cytolytic T-lymphocyte (CTL)’ effec-
tor functions is the exocytosis of cytolytic granules (22, 23).
However, in our studies of molecular mechanisms of T-cell
mediated cytotoxicity we found serious inconsistencies with
this model (24). It was then demonstrated by us (25) and
others (26) that in the absence of extracellular Ca*‘, exocy-
tosis of cytolytic granules that contain pore-forming protein
is completely blocked, while significant CTL mediated lysis
of target cells still occurs. Thus, we searched for alternative
mechanisms utilizing a lytic intermediate that can kill target
cells even in the absence of extracellular Ca*+.
The cell membrane permeabilizing properties of ATP and
its profound effects on cell physiology (7-21) suggested the
distinct possibility that extracellular ATP may be involved in
mechanisms of cell-mediated cytotoxicity. However, we could
not find in the literature the description of cytolytic properties
of extracellular ATP. Since the important assumption of this
hypothesis is the ability of ATP to cause cell death, we
investigated the fate of cells after 2-6-h incubations with
ATP.
During the course of these studies we indeed found that
ATP can kill different cells, except cells that express a high
level of ATP-degrading activity on their surface. Assuming
that extracellular ATP could serve as a mediator of CTL
effector functions, we expected CTL themselves to be highly
resistent to effects of ATP. We found that CTL do indeed
express a high level of ecto-ATPase activity, which protects
them from the lytic effects of extracellular ATP. In contrast
to other well studied ATPases involved in ion transport,
muscle contraction, non-muscle cell mechanical activities,
and energy transduction ecto-ATPases only recently attracted
significant attention. In this paper we describe the catalytic
properties of ecto-ATPase in lymphocytes, experimental ma-
nipulations that modulate its activity, and its possible biolog-
ical role.
EXPERIMENTAL PROCEDURES
Materials
ATP (disodium salt from equine muscle), ATP (prepared by phos-
phorylation of adenosine), AMP-PNP, EGTA, EDTA, FSBA, ADP,
AMP, GTP, UTP, ITP, oligomycin, and Hepes were obtained from
Sigma. Luciferin, luciferase, and ATPys were purchased from Boeh-
’ The abbreviations used are: CTL, cytolytic T-lymphocytes; AMP-
PNP, 5’-adenylyl-P,r-imidodiphosphate; ATP-yS, adenosine B’-thio-
triphosphate; EGTA, Iethylenebis(oxyethylenenitriio)]tetraacetic
acid; FiTC, fluorescein 6’-iskhiocyanaie; FSBA, 5’-fluorosulfonyl-
benzoyladenosine; PEL, peritoneal exudate CTL; Hepes, 4-(2-hy-
droxyethyl)-1-piperazineethanesulfonic acid; LDH, lactate dehydro-
genase.
334
by guest, on July 10, 2011www.jbc.orgDownloaded from
Ecto-ATPase in Cytotoxic T-lymphocytes
ringer Mannheim. [Y-~~P]ATP and [T-~‘P]GTP were purchased from
New England Nuclear (Pittsburgh, PA). values for triplicate cultures were determined. Percentage of specific
61Cr release was calculated as 100 X (a -
b)/(t - b).
where
a
is “Cr
release in the presence of nucleotide,
b
is spontaneous release from
labeled cells, and
t
is the total release determined from supernatants
of cells lysed by 0.1% Triton X-100.
Cell Culture and Maintenance
CTL clones 2C (30) and OE4 (31) are normal, antigen and inter-
leukin-2-dependent cell lines which are heavily used in the
in vitro
studies of effector functions of lymphocytes. They were maintained
as described previously (24, 30, 31). CTL were purified from dead
cells by Ficoll-Hypaque centrifugation shortly before being used in
assays. P815 mastocytoma cell line, P815.2 subline, and EL4 tumor
cell line were used as target cells and were maintained in RPM1 1640
(Biofluids, Rockville, MD) containing 10% fetal calf serum.
Peritoneal exudate CTL (PEL, alloimmune peritoneal exudate
lymphocytes with BALB/c anti-EL4 specificity) were obtained 5 days
after secondary intraperitoneal immunization of BALB/c mice with
25 X lo6 EL4 tumor cells (prepared as ascites in syngeneic C57BL/6
mice) (32). PEL were purified from crude peritoneal exudate cells by
removal of macrophages and other adherent cells on nylon wool
columns (60 min at 37 “C) followed by elution of the nonadherent
(PEL).
Assay of ATPase Activity in Cell Suspensions, Homogenates, and Cell
Extracts
Assay
of Enzymatic Activity Using
[T-~‘P]ATP--T~~ membrane-
bound ATPase activity was measured by using ~Y-~*P]ATP as sub-
- _.
strate by counting the-amount of liberated Pi (inorganic phosphate)
in supernatants after precipitation of unhydrolyzed [y-32P]ATP with
activated charcoal (27). The standard reaction medium in a final
volume of 0.2 ml contained 10 mM Hepes (pH 7.4), 3 mM ATP (0.3
PCi of [T-~‘P]ATP), 135 mM NaCl, 5 mM KCl, 2 mM Ca&, 2 mM
MgCl*, 10 mM glucose, 1% bovine serum albumin, and 5 X lo4 CTL.
After 10 min in 37 “C water bath the reaction was stopped by adding
0.5 ml of cold 20% (w/v) activated charcoal in 1.0 M HCl. The tubes
were cooled on ice fo; 10 min and centrifuged at 3200 X g at 4 “C.
Aliquots (0.2 ml) of the clear supernatant were transferred to scintil-
lation vials containing 2.5 ml of Readv-Gel scintillation mixture and
counted in a Beckman counter for “P;.
Molybdate
Assay
of ATPase Activity-CTL
were purified by Ficoll-
Hypaque and washed twice with buffer A which contained 120 mM
NaCl, 5 mM KCl, 20 mM Hepes/Tris (pH 7.4). CTL (0.5 X lo6 cells)
in buffer A were added to tubes containing Ca*+ solution to give a
final concentration of 2 mM (for Ca*‘-ATPase activity) or Mg2
solution to give a final concentration of 2 mM (for Mg2’-ATPase
activity) in total volume 400 ~1. After incubation of cells at 37 “C for
5 min, the assay was started by addition of ATP (or another nucleotide
triphosphate) to a final concentration of 2 mM. After incubation the
reaction was stopped by adding 100 11 of sodium dodecyl sulfate to a
final concentration of 2%. The ATPase activity was determined by
measuring the inorganic phosphate released as described by Ames
(28). Briefly, 0.3 ml of the phosphate solution (or 0.3 ml of buffer
containing 120 mM NaCl, 5 mM KCl, 20 mM Hepes/Tris (pH 7.4) for
the blank) was added to 0.7 ml of a mixture composed of 1 part of
10% ascorbic acid, plus 6 parts of 0.42% ammonium molybdate in 1
N H2SOd, and incubated for 1 h at 37 “C. The absorbance of samples
at 820 nm was measured using Beckman DU-50 Spectrophotometer.
Luciferin-Luciferase Assay of Ecto-ATPase
Activity-CTL were
purified and washed twice with M-Krebs solution (135 mM NaCl, 5
mM KCl, 2 mM CaCl2, 2 mM MgC12, 10 mM glucose, 1% bovine serum
albumin, 10 mM Hepes at pH 7.6). CTL (3 X lo5 cells) were mixed in
the luminometer cuvette with 100 ~1 of a solution composed of
luciferin (0.28 mg/ml) and luciferase (0.0016 mg/ml) in M-Krebs and
100 ~1 of a solution of 60 nM ATP in a final volume of 300 pl. With
the luciferin-luciferase assay the luminescence intensity in the cuvette
is proportional to the concentration of ATP. Decrease in intensity in
luminescence was used as a measure of ATP-degrading activity in
the present studies. An LKB 1250 luminometer was used.
Cytotoxicity Assays
Before addition of ATP, ATPrS, or AMP cells were labeled with
‘ICr by incubating with 100 &i of Na,CrO, for 1 h, 37 “C in CO,
incubator followed by two washings and Ficoll-Hvpaque purification
of intact cells. ‘ICr release assay-was performed-essentially as de-
scribed (25, 29). Briefly, 51Cr-labeled tumor and CTL clone cells were
cocultured in triplicate in a 96-well plate. Different nucleotides were
added at a final concentration ranging from 0.5 to 8 mM. After
incubation the 96-well plate was centrifuged (1 min, 800 rpm, 4 “C),
radioactivity in culture supernatants was measured, and the mean
Lactate dehydrogenase (LDH) activity was assayed as we described
earlier (24). The activity of LDH, a marker enzyme of the cytosolic
compartment, was tested in supernatants of tumor cells or CTL after
4-h incubation as an indicator of cell death.
The cell death was checked by trypan blue stain. The results
correlate well with LDH assay and ‘ICr assay.
“Rescue” Assay of Ecto-ATPase Activity
This assay is based on the assumption that a decrease in extracel-
lular ATP will be reflected in decreased ATP-induced death of
“detector” target cell. Thus, the presence of ecto-ATPase activity on
the surface of CTL introduced into an assav of ATP-induced ‘ICr
release from target cells not recognized by th”e CTL clone will result
in decreased target cell death during incubation with ATP. EL4 target
cells were chosen for this experiment because CTL OE4 do not
recognize and do not kill them;therefore, observed ‘ICr release is due
only to ATP-induced EL4 destruction, but not to CTL-induced EL4
death. The ability of ATP to kill EL4 tumor cells was well character-
ized in preliminary experiments. ‘ICr-labeled cells were cultured for
4-6 h with ATP (at concentration that causes significant cell death)
in the presence or absence of different number of CTL OE4 cells.
Data are presented as ATP-induced ‘ICr release from EL4 cells
dependent on the cell ratio of OE4:EL4. where the ratio “0” corre-
spbnds to EL4 in the presence of ATP and in the absence of CTL
OE4.
Pretreatment of CTL with FITC
and
FSBA
Clone OE4 CTL at lo6 cells/ml were incubated in RPM1 media at
0, 20, and 37 “C with different concentrations of FITC or FSBA in
the presence or absence of 2 mM ATP. After incubation, excess
reagent was removed by washing cells in M-Krebs solution. Activity
of ecto-ATPase and functional activity of cells was tested as described
above.
Substrate specificity experiments
were performed by incubating
cells with [T-~*P]ATP or [-y-3*P]GTP in the presence or absence of
increasing concentrations of other unlabeled nucleotide triphosphates
and measuring the liberation of 32Pi using the charcoal method as
described in the figure legend or by comparing the rates of release of
Pi using molybdate assay after incubation of different unlabeled
nucleotide triphosphates with the same number of cells.
RESULTS
Cytolytic Properties of Extracellular
ATP--Incubation of
cultured cell lines with ATP or ATP analog (ATPTS) (34)
resulted in the destruction (51Cr release) of EL4 thymoma
cells (Fig. lA) and P815 mastocytoma cells (Fig. lB), but not
of CTL clone (Fig. lc), even when the concentration of ATP
was raised to 8 mM. AMP had no effect at any tested concen-
tration on any cell line. Significant cytotoxicity was observed
even at 0.5 mM ATP with EL4 cells, but not with P815 cells.
Different ATP preparations were tested, including “cell cul-
ture grade” (Sigma, Cat. No. A3284), ATP from equine muscle
(Sigma, Cat. No. A5394), and synthetic ATP preparations
(Sigma, Cat. No. A3377) with the same effect on viability of
EL4 or P815 cells. It is remarkable that CTL clone was not
lysed by ATP at concentrations that caused a high degree of
EL4 or P815 cell death. The same results were obtained using
trypan blue exclusion test and LDH release assay, which were
used independently to confirm effects of extracellular ATP.
For example, 1 and 5 mM ATP caused 8 and 28% specific
LDH release from EL4 cells after incubation for 4 h.
Time course of ATP- or ATPyS-induced cell lysis (data
not shown) is similar to that described for CTL-induced 51Cr
release (33), since significant EL4 and P815 lysis was observed
already after 1 h, and practically all cells were dead after 7 h.
In some experiments ATPyS was almost as efficient as ATP
in killing EL4 cells, but was always two to three times less
lytic toward the P815 cells than ATP (Fig. 1,
A
and B).
by guest, on July 10, 2011www.jbc.orgDownloaded from
336 Ecto-ATPase in Cytotoric T-lymphocytes
FIG. 1. Effect of ATP, ATPrS, and AMP on the 61Cr release from normal T-lymphocyte, thymoma,
and mastocytoma cell lines. After “‘Cr-labeling cytolytic T-lymphocyte (CTL clone OE4), EL4 (thymoma cell
line), or P815 (mastocytoma cell line) were incubated in CO, incubator at 1 X lo6 cell/ml in RPMI, 5% fetal calf
serum for 4 h with or without adenine nucleotides. Percent of specific ‘ICr release was calculated as described
under “Experimental Procedures.” A, effect on 51Cr release from EL4 cells. B, effect on Wr release from P815
cells. C, effect on ‘Cr release from CTL OE4 cells.
ATP IOmMtZmMEGTA
ATP 1 OmM
ATP 5mM+2mMEGTA
w EL4
40
+ 0!34:EL4+ATPhM
x * OE4:EL4
z
30
* OE4 : EL4 + ATFpmM
f
ATP 5mM
ATP 1 mM+ZmMEGTA
ATP 1mM
-20 0 20 40 SO 90 100
% specific SlCr-release
FIG.
2. Lysis of cultured cells by ATP in the presence and
absence of Ca”+ in the incubation media. “Cr-labeled EL4, P815
cells, or P815.2 (Ca’+-insensitive subline) cells were incubated with 1
mM
+ 10
mM
ATP in the presence or absence of 2
mM
EGTA for 4
h, 37 “C in CO2 incubator, and percent of specific ‘ICr release was
calculated as under “Experimental Procedures.”
Among other adenine nucleotides tested only ADP had any
measureable lytic effect toward EL4, although always less
than ATP. ADP had no effect on P&315 cells, and AMP had
no effect in any tested cellular system (Fig. 1, A-C).
Lytic effects of ATP were not affected by the addition in
the incubation medium of EGTA to chelate the extracellular
Cal+ (Fig. 2). Both sublines of P815 (P815 and P815.2) as
well as EL4 cells released approximately the same proportion
of intracellular radiolabel with or without EGTA in the in-
cubation medium.
ATP-degrading Activity on the CTL Surface-The resist-
ance of CTL clone to high concentrations of extracellular
ATP suggested that it could be due partly to their increased
cell surface ATP-degrading activity, possibly mediated by the
expression of high amounts of ecto-ATPase. The experiments
were therefore designed to test in a biological assay if CTL
do indeed have this activity, thus lowering extracellular con-
centration of this lytic agent. It was expected that lytic
potential of extracellular ATP toward the ATP-sensitive cells
would be greatly diminished in the presence of CTL. Such
experiment could be performed using a CTL clone and 5’Cr-
labeled cells that are not recognized by that clone, so that
ATP-induced “‘Cr release from labeled target cell would be
distinguised from CTL-induced ‘Cr release. CTL clone OE4
and EL4 cells provide a convenient combination, since these
0:1
1 :1 3:1 9:t 27:l
CTL OE4: EL4 target cell ratio
FIG.
3. Protection of EL4 cells from lysis by 5
mM
ATP with
CTL OE4 cells. Rescue assay. Wr-labeled EL4 cells were incubated
with 5 mM ATP or ATPrS with varying numbers of CTL clone OE4
cells. After 4 h ‘ICr release was calculated. N, Wr release induced by
5
mM
ATP;
A,
Wr release induced by 5
mM
ATPrS; 0, 51Cr release
of EL4 incubated with different numbers of CTL OE4 in the absence
of ATP or ATPrS.
target cells are not recognized by CTL OE4 (31), and EL4
targets are routinely used in our studies as a negative control
for the antigen-specificity of CTL OE4. CTL OE4 exquisitely
recognizes and kills only H-2d antigen-bearing target cells,
whereas EL4 expresses H-2b. According to the design of this
experiment CTL are present in the assay only as a source of
the ATP-degrading activity, but not in their capacity to kill
target cells, causing 51Cr release. Thus, evaluation of the
experiment that is presented in Fig. 3 is based on the detection
of 51Cr release from radiolabeled EL4 cells that are incubated
with 5 mM ATP or ATP$S, in the presence of varying
numbers of CTL OE4. Inhibition of such ATP-induced ‘ICr
release from EL4 by CTL is suggestive evidence for ATP-
degrading activity on the surface of CTL. Fig. 3 demonstrates
that CTL can protect (rescue) EL4 cells from lysis by ATP,
since much less “‘Cr release was detected when CTL OE4
were present during incubation of Wr-labeled EL4 with ATP.
It is shown in Fig. 3 that protection from ATP-induced cell
death is dependent on the number of CTL present. The
interesting dichotomy is observed in the ability of CTL to
rescue EL4 cells from cytotoxic effects of ATP and ATPyS
(Fig. 3). In this particular experiment 5 mM ATP caused 38%
specific 51Cr release, whereas 5 mM ATP-&l induced only
by guest, on July 10, 2011www.jbc.orgDownloaded from
Ecto-ATPase in Cytotoxic T-lymphocytes 337
about 15% 51Cr release. However, although addition of in-
creasing number of CTL per well (to a maximal ratio of 27
CTL/target cell) dramatically decreased ATP-induced 51Cr
release, it did not affect ATPyS-induced 51Cr release. Al-
though both ATP and ATPyS can be utilized by protein
kinases, dephosphorylating enzymes are much less efficient
with ATPyS-phosphorylated proteins (34). It also would be
expected that ecto-ATPase will be less efficient in hydrolyzing
ATPyS than ATP. It appears, that the rescue of EL4 from
lysis by CTL is most likely due to ATP-degrading activity of
CTL (ecto-ATPase).
Measurements and Biochemical Characterization
of
Ecto-
ATPase Activity in Intact
CTL-Since ecto-ATPase activity
is measured by incubation of cells with ATP, and since we
demonstrated that extracellular ATP can kill cells, causing
release of intracellular enzymes and possibly intracellular
ATPase, considerable effort was made to localize the meas-
ured activity to a plasma membrane ecto-enzyme with its
nucleotide hydrolyzing site facing into extracellular space.
Luminometry provides a sensitive measure of ATP concen-
tration using a luciferin-luciferase assay system (see “Exper-
imental Procedures”) and can be applied to the determination
of ecto-ATPase activity on CTL by addition of these cells
into the reaction mixture.
Degradation of ATP was evident in the first seconds after
addition of CTL into the luminometer chamber (Fig. 4A).
Some spontaneous slow decrease in intensity is observed;
however, addition of 3 x lo5 CTL OE4/300 ~1 of incubation
mixture results in the sharp decrease of luminescence, reflect-
ing the ATP degradation by CTL ecto-enzymes. In contrast
to CTL OE4, very little change in luminescence intensity
(ATP degradation) could be detected when tumor cells EL4
and P815 were added into the luminometer. Such time course
is not consistent with the alternative explanation, that leak-
age of intracellular ATPases accounts for degradation of
extracellular ATP. With 1 mM ATP (the concentration used
in ATPase assay) no LDH release from CTL OE4 was de-
tected at all (data not shown). In addition, no damage to CTL
during first 30 min of incubation in ATPase assay conditions
was detected by trypan blue exclusion test. Additional evi-
dence against the possibility that the cell damage and conse-
quent release of intracellular ATPases account for the meas-
ured ecto-ATPase activity comes from the studies of ecto-
ATPase activity where non-ionic detergent was used in the
I
I I
I I
CTL, OE4 I
E\4 I p815
FIG.
4. Demonstration of ecto-
ATPase activity in lymphocytes by
luciferin-luciferase assay. A,
inten-
sity of luminescence was monitored dur-
ing luciferin-luciferase reaction that uti-
lizes ATP after the addition of equal
number (3
x
105 cells) of CTL OE4, EL4,
and P815. B, comparison of ecto-ATPase
activity in intact cells and Triton X-100
solubilized cells (lo5 cells).
t
ATP 2&M
TX-100
ATP MnM
TIME
by guest, on July 10, 2011www.jbc.orgDownloaded from
338 Ecto-ATPase in Cytotoric T-lymphocytes
course of luciferin-luciferase assay (Fig. 4B). When CTL
plasma membranes are purposefully damaged by the addition
of non-ionic detergent, this results in immediate release of
intracellular ATP (sharp increase of luminescence indicated
by the arrow on the figure). However, no significant change
in general rate of ATP hydrolysis was detected. Thus, even if
cells are broken, it does not contribute significantly to detect-
able ecto-ATPase activity. In separate control experiment we
found that ATP, present during the assay of ecto-ATPase in
CTL is not likely to cause cell damage leading to release of
intracellular ATPases. Only 5% of total LDH could be de-
tected in supernatants of intact CTL after incubation for 30
min with the concentration of ATP used to detect ecto-
ATPase activity in 32Pi-charcoal assay, whereas practically all
of cellular LDH were found in the broken cells homogenates.
Effect of Ca2+ and Mg2+ on Ecto-ATPase Actiuity-Ecto-
ATPase in CTL is Ca2+ and Me-dependent as demonstrated
in Fig. 5. No increase of total activity of ecto-ATPase was
observed, when both Mg2+ and Ca2+ were present during assay
of intact cells (Fig. 5). The data indicate that both activities
are produced by the same ecto-enzyme, since additive effects
would be expected if there were separate Ca”-dependent and
Me-dependent enzymes. In the absence of added Ca2+, the
ecto-ATPase activity of CTL is stimulated by MC, whereas
in the absence of free Mg2+, ecto-ATPase activity is stimulated
by Ca2+.
Substrate Specificity of Ecto-ATPase Activity in
CTL-The
nucleotide specificity was determined by measuring inorganic
phosphate (28). Ecto-ATPase in CTL has broad substrate
specificity (Table I), since GTP, UTP, and ITP were hydrol-
ized, even at higher rate than ATP, whereas AMP and AMP-
PNP were not hydrolyzed. ADP was hydrolyzed at the rate of
86 and 73% of ATP with Ca’+-dependent and Mg2’-dependent
activities, respectively. It is possible, however, that there is
more than one nucleotide 5’-triphosphatase on the surface of
CTL. The assay based on the measurements of inorganic
phosphate release from unlabeled ATP nucleotide triphos-
phate does not distinguish such activities.
Therefore, we utilized [-y-32P]ATP charcoal assay to com-
pare effects of different nucleotides on hydrolysis of ATP
(Fig.
6A).
It is found that ATP and ADP were most efficient
in blocking [-y-32P]ATP hydrolysis, with 50% inhibition at
approximately 0.025 and 0.5
mM,
respectively, and complete
inhibition at 2-4
mM.
GTP, ITP, and UTP were inefficient.
FIG.
5. Ca’+ and Mg2+ dependence of ecto-ATPase activity
of intact cells and of ATPase activity of the same number of
homogenized cells. Ficoll-purified, intact CTL OE4 (viability above
98%) (5 X 10’ cells/tube) were incubated for different times with [y-
32P]ATP and the amount of liberated 32Pi was estimated using char-
coal assay as described under “Experimental Procedures.”
n
, ecto-
ATPase activity in the presence of EDTA (2
mM);
q
a, 2 mM Ca*+
were
added;
q
(2
mM
Mg2+ were added; EZl , both 2
mM
Ca2+ and Mg2
were present during the assay.
TABLE I
Substrate
specificity of ecto-Ca2+
-ATPase and Mg”-ATPase on CTL
OE4
Activity of ecto-ATPase of CTL OE4 was tested using molybdate-
based assay for the detection of inorganic phosphate, as described
under “Experimental Procedures.” The specific activities for 100%
Mg2+-ATPase and 100% Ca’+-ATPase (in parallel Pi) were: 0.76
nM/
min/105 cells for Mp-ATPase and 0.64 nM/min/lO’ cells for Ca’+-
ATPase as calculated according to Ref. 28.
Substrate
ATP
ADP
AMP
AMP-PNP
GTP
UTP
ITP
W+-ATPase Mg2’-ATPase
%
100 100
86 73
0.2 0.3
0 0
119 118
200 173
144 160
1000ol
A
.s 6000
ATP
ADP
GTP
ITP
UTP
Nucleotide added (mM)
- ATP
- GTP
--z- ITP
-fr- UTP
.v
:
i loooo
.-
i
s
?
v
0 0.0 05 1.0 I.5 2.0 2.5 3 0 3.5 4 0 4.5
Nucleofide added (mM)
FIG.
6. Substrate specificity studies of ecto-ATPases in
CTL. CTL OE4 (5 X lo4 cells/ml) were incubated with 0.5 pCi of
[r-s’P]ATP (A) or 0.2 &i of [T-~*P]GTP (B) for 10 min, 37 “C in
the presence of different concentrations of other adenine nucleotides.
After incubation, the amount of liberated 32Pi was estimated. The
total concentration of ATP (A) and GTP (B) in the assay media was
200
WM.
Since GTP, UTP, and ITP are hydrolyzed by CTL ecto-
enzymes at higher rates than ATP (Table I), there is appar-
ently
more than one enzyme with nucleoside 5’-triphospha-
tase activity. If Ca2+/Mgs+-dependent ecto-ATPase and Ca2+/
Mg2+-dependent GTPase have different substrate specifici-
ties, then
it would be expected that GTP is efficiently hydro-
lyzed by ecto-GTPase, but is not efficient in blocking the
hydrolysis of ATP by ecto-ATPase. This possibility was tested
in [T-~‘P]GTP hydrolysis assay by studying the ecto-GTPase
activity of intact CTL OE4 in the presence of different con-
centrations of other nucleotide triphosphates (Fig. 6B). Sur-
prisingly, GTP was no more efficient in competing with [y-
32P]GTP than ATP or ADP and ITP. Therefore, it appears
by guest, on July 10, 2011www.jbc.orgDownloaded from
Ecto-ATPase in Cytotoxic T-lymphocytes 339
that ecto-ATPase activity on the CTL surface is represented
by more than one Ca2+/Mg2+-dependent enzyme; a Ca*+/Mg’+
ecto-ATPase with higher selectivity toward the ATP and ADP
and another ecto-nucleoside triphosphatase activity with a
broad substrate specificity (Fig. 6B).
Effect of Inhibitors on E&o-ATPase Actiuity-FITC and
FSBA have been used before (35) for the affinity labeling of
the ATP-binding site of various enzymes. We made an as-
sumption that ecto-ATPase may contain one or more ATP-
binding sites reactive with these covalent affinity labels. The
5’-p-fluorosulfonylbenzoyladenosine (5’-FSBA) is a synthetic
adenosine analog, which is capable of covalent reaction (al-
kylation) with proteins. It was shown that FSBA labels two
different sites on the subunit of the dog kidney Na+/K+-
transporting ATPase, and amino acid sequences of the labeled
tryptic peptides were described (35).
We found that FSBA strongly inhibits ecto-ATPase activity
of cloned CTL (Fig. 7), whereas FITC and vanadate pretreat-
ments have no effect (data not shown). Addition of ATP
together with FSBA during CTL pretreatment completely
protected ecto-ATPase activity from FSBA inactivation, in-
dicating that the molecular target of FSBA is an extracellular
ATP-binding site on CTL. The same results were obtained
when ecto-ATPase activity and susceptibility of ecto-ATPase
to inhibition by FSBA were tested in polyclonal CTL obtained
from PEL (data not shown). When time dependence of pre-
treatment with FSBA was studied, it was found that 5 min of
pretreatment at room temperature is sufficient to cause 44%
inhibition of ecto-ATPase activity. No dramatic differences
in effect of temperature on effect of FSBA were noted. Pre-
treatment of CTL for 30 min at 0, 20, and 37 “C resulted in
59,78, and 85% inhibition, respectively. These data suggested
that it is possible to pretreat cells with FSBA at 0 “C, thereby
minimizing possible and unwanted interactions of FSBA with
ATP-binding sites of intracellular proteins. Such an experi-
ment is presented in Fig. 8. It is shown that even 10
mM
ATP
is not able to kill CTL clone when ecto-ATPase is intact.
However, when CTL ecto-ATPase is inactivated by pretreat-
ment with FSBA, CTL become susceptible to lytic effects of
ATP. When 3
mM
ATP was present during pretreatment of
CTL with FSBA (thus protecting ATP-binding sites on CTL),
120
.r
s
100
W FSBA
:
q
FSBA+ATP
2 60
5
60
.z
.-
: 40
::
s- 20
l-l
1 10 100 1000
,m of FSDA
FIG. 7.
Inhibition of ecto-ATPase activity of CTL clone by
FSBA. lo6 CTL OE4 or PEL were preincubated for 40 min with 1,
10,100, or 1000
FM
of FSBA or with the same concentration of FSBA
but in the presence of 3
mM
ATP, and excess of FSBA was washed
away. Effect of FSBA was calculated as a percent of specific ATP
hydrolysis, where rate of hydrolysis of ATP in the presence of 0.5%
dimethyl sulfoxide (v/v) is considered as 100%. FSBA was added as
solution in dimethyl sulfoxide. Maximal addition of dimethyl sulfox-
ide did not exceed 0.5%. In parallel experiments 3
mM
ATP were
added together with FSBA to demonstrate the extracellular location
of ATP-binding site that is modified by FSBA.
FSBA+ATP
I . I . I . I - I . I . I
-10 0 10
20 30 40 50
56 control specific 510.release
FIG.
8. Inactivation of ecto-ATPase on CTL by pretreat-
ment with FSBA results in the loss of resistance of CTL to
lytic effects of extracellular ATP. Wr-labeled cells (CTL OE4
at 106/ml) were preincubated with FSBA (1
mM)
as described under
“Experimental Procedures” in the presence or absence of 3
mM
ATP.
After the washing control, FSBA-pretreated and (FS’EA + A!FP)
pretreated CTL were incubated additional 4 h at 37 “C in the presence
of 10
mM
ATP and percent of specific “Cr release was estimated.
no ecto-ATPase inactivation was observed (Fig. 7) and CTL
were still resistant to the lytic effects of ATP.
DISCUSSION
Some of the properties of the enzyme activity described
here resemble ecto-ATPase, described in Erlich ascites car-
cinoma cells (36), since ATP, GTP, UTP, CTP, and ITP at
1 mM concentration, as well as ADP and AMP, are hydrolyzed
at similar rates. However, AMP is not hydrolyzed at all by
CTL ecto-ATPase, whereas carcinoma cells do hydrolyze
AMP at rate comparable to ATP (36). In contrast to another
described ecto-ATPase in human blood platelets (37), where
GTP and UTP are hydrolyzed at the rate of 7 and 13% the
rate of hydrolysis of ATP, respectively, CTL ecto-ATPase
hydrolyzes GTP at 119% and UTP at 200% of the rate of
hydrolysis of ATP.
Our substrate competition studies suggest that there is at
least one nucleoside 5’-triphosphatase with broad nucleotide
specificity and an ecto-ATPase that preferentially hydrolyzes
ATP. The murine lymphocyte ecto-ATPase described here as
having broad specificity, hydrolyzes all nucleoside 5’-triphos-
phates with similar rates (Fig. 6B and Table I). Therefore,
this enzyme could be called ecto-nucleoside triphosphate
phosphohydrolase, resembling one described by DePierre and
Karnovsky (38) in guinea pig polymorphonuclear leukocytes
and by Lin and Russel (39) in hepatic cells. However, in that
study competition experiments were not performed. That
enzyme was also not inhibited by oligomycin nor stimulated
by DNP, confirming that ecto-ATPase is not from mitochon-
dria.
To avoid the artifact common with ecto-enzymes (40), we
made a concerted effort to assure that cell preparations con-
tained virtually
100%
intact cells. That was achieved by
carefully cultivating cells in the optimal conditions, washing,
and purifying live cells by Ficoll centrifugation immediately
before use in experiments. The strongest evidence for the
surface location of ATPase is provided by ATPase activity
measured by luminometry (Fig.
4),
where this activity was
detected in the first seconds after cell addition without any
noticeable loss of viability of cells. Even when cells were
partially broken by low concentrations of non-ionic detergent
by guest, on July 10, 2011www.jbc.orgDownloaded from
340 Ecto-ATPase in Cytotoxic T-lymphocytes
that result in release of intracellular ATP, we did not detect
increase in the rate of ATP hydrolysis (Fig. 4). Additional
evidence for extracellular location of studied enzymatic ac-
tivities are provided by: (i) the linearity of ecto-ATPase
activity in time course studies, that reflects absence of the
leakage of the cell membranes; (ii) the lack of release of
intracellular marker enzymes, including both cytoplasmic
(LDH) and granular (BLT-E) enzymes, tested in parallel
experiments to assure that no cell leakage takes place during
the assay; (iii) the ability of CTL to protect another cell from
lytic effects of ATP. Finally, the ability of extracellular ATP
to protect CTL ecto-ATPase from inactivation by FSBA
(Figs. 7 and 8) provides convincing evidence for the location
of ATPase on the cell surface.
Ecto-ATPases are enzymes in search of a function, since
no evidence has been yet presented on their role in specific
cell-mediated effects. It was suggested that possible role of
ecto-ATPase is to terminate the effect of ATP on the cells or
to regulate effects of extracellular ATP due to its phosphatase
activity (41).
Different conditions of incubation and concentration were
tested in an attempt to find a specific inhibitor of ecto-
ATPase. Neither FITC nor vanadate inhibited this enzymatic
activity in CTL, but FSBA was found to be very efficient,
indicating the usefulness of FSBA in future studies of the
functional role of ecto-ATPase in lymphocytes. FSBA is
especially preferred in situations where internal controls are
needed to specifically target cell surface enzyme using a
covalent affinity label. The use of FSBA, however, is limited
by the distinct possibility of its interaction with ATP-binding
sites of other enzymes. Therefore, the work is now in progress
to develop a monoclonal Ab against CTL ecto-ATPase. Of
several assays of ecto-ATPase activity that were used here,
the rescue assay holds promise as a screening assay in the
development of monoclonal Ab against surface ATPase. We
are screening hybridomas for the ability to block CTL-me-
diated rescue of the ‘Q-labeled EL4 cells, incubated with
lytic concentrations of extracellular ATP.
Our results suggest a functional role of ecto-ATPase, which
was not discussed previously: protection of effector cells of
the immune system in areas where local concentrations of
extracellular ATP could be quite high due to specific release
by effector cells or to cell death. When cells are damaged by
physical injury or by other means that are not accompanied
by extensive degradation, the neighbors of damaged cells may
be transiently exposed to concentrations of ATP in the mil-
limolar range (6). It is also an attractive hypothesis that ecto-
ATPase of CTL protects them from the lytic intermediate
(ATP) that is involved in the delivery of the lethal hit to the
target cell. At this point the working hypothesis is that
extracellular ATP plays an important role in CTL functions
by being utilized by ecto-protein kinases. The new model of
cellular cytotoxicity we test now is based on assumptions that
CTL can secrete ATP +oward the Ag-bearing target cell in
the CTL-target c0njuga.a. We speculate that ATP in concert
with another surface molecule affects the properties of the
target cell membrane and intracellular metabolism resulting
in target cell death. The preliminary results are in agreement
with these key assumptions.* Especially relevant is the most
recent observation of secretion of ATP by receptor-stimulated
CTL and PEL. The role of CTL ecto-ATPase according to
this hypothesis is to protect CTL from lytic effects of ATP
which could be accumulated in high local concentrations in
the area of contact between CTL and target cells. We started
our studies of ecto-ATPase in CTL in an attempt to explain
the mechanism of self-protection of CTL during CTL-me-
diated target cell lysis in the presence of 2
mM
MgEGTA (24,
25). It is an important property of this enzyme in CTL that
it is fully active in these conditions, since the presence of
M$+ is sufficient. Thus, biochemical properties of CTL’s
ecto-ATPase are consistent with the functions postulated.
AcknowZedgment-The editorial assistance of Shirley Starnes is
greatly appreciated.
REFERENCES
1. H&on, P. (1959)
J. Ph siol.
146,494~504
2. Katsuragi, T., and Furu
H
awa, T. (1985)
Trends Phormocol. Sci. 6,337-339
3. Danzi er R. S., Raffaeli S., Horeno-Sanchez R. Sakai H. Capo ossi,
M. 8. ‘Sporgeon H. A:, Hansford, R. G., and Lakatta, E. 6. (19& Cell
Cal&m 9,193-i99
4. Haslam, R., and Cusack, N. (1981)
Purinergic Receptors,
Chapman and
Hall, London
5. Gordon. E. L.. Pearson. J. D.. and Slakev. L. L. (1986)
J. BioL Chem.
261,
154SR-15xi4
_ _ _ _ - _ _
6. Holmsen, H. (1985)
Annu. Reu. Physiol. 47,677-690
7. Gordon, J. I. (1986)
Biochem. J. 233
309-319
8. Bennett, J. S., Colman, R. F., and dolman, R. W. (1978)
J. BioL Chem.
263,7346-7354
9. MacManus, J. P., and Whitfield, J. F. (1972) Life
Sci.
11, 837-845
10. Evans, R. H. (1974) Nature 250,391-394
11. Gregory, S., and Kern, M. (1978) Biochem.
Biophys. Res. Commun. 63,
1111-1116
12. Gregory, S. M., and Kern, M. (1981)
Immunology 42,451-457
13. Diamant, B., and Kruger, P. G. (1967)
Acta Physiol. Stand.
21,291-295
14.
Lisifii Knshnaraj, R., and Kemp, R. G. (1985) J. Immunol.
136, 3403-
15. Okajima, F., Tokumitsu, Y., Kondo, Y., and Ui, M. (1987)
J. BioL Chem.
262,13483-13490
16. Steinberg, T. H., and Silverstein, S. C. (1987)
J. BioL Chem. 262,
3118-
“tn.7
01“
17. Steinhere. T. H.. Newman. A. S.. Swanson. J. A.. and Silverstein. S. C.
18.
19.
20.
21.
(1987)“J. Biol. them. 26&88&i-8888 ’
Buisman H. P., Steinberg, T. H., Fischbourg, J. Silverstein, S. C., Vo el-
sang, S. A., Ince, C., Ypey, D. L., and Leigh, P. C. J. (1988)
Proc. h&l.
Acad. Sci. U. S. A. S&7988-7998
Greenberg, S., Di Virgilio, F., Steinberg, T. H., and Silverstein, S. C. (1988)
J. BioL Chem. 263,10337-10343
Cockroft, S., and Gomperts, B. D. (1980)
Biochem. J.
186.789-798
Tatham. P. E. R., Cusack, N. J., and Gomperts, B. D. (1988)
Eur. J.
Pharmacol.
147. 13-21
Henkart, P. A. (1985)
Annu. Rev. ImmurwL 3,31-58
Podack, E. R., and Konigsberg, P. J. (1984)
J. Ex Med.
160,695-710
Takayama, H., and Sitkovsky M. (1987)
J. Ex
Trenn, G., Taka
Ostergaard, H. E ama, H., and Sitkovsky, M.
q ffed.
166,725-743
. (1987) Nature 330,72-74
., Kane, K. P., Mesher, M. F., and Clark, W. R. (1987)
Nature 330,71-72
Reddy, N. B., Oliver, K. L., Festoff, B. W., and Engel, W. K. (1978)
Biochem. Bio hys. Acta 640,371-388
28. Ames, B. N. (1 66)
Methods Enz
8 mol. 8.115-117
29. Grabstein, K. (1980)
Selected
A!?
ethods in Cellular Immunology,
pp. 124-
137 W. H. Freeman & Co., San Francisco
30. Berrebi, G., Takayama, H., and Sitkovsky, M. W. (1987)
Proc. NatL Acad.
Sci. U. S. A. 84,1364-1368
31. Staerz, U. D., Kanagawa, O., and Bevan, M. J. (1985)
Nature
314, 628-
fill
“YL
32. Be;[iiG., Sullivan, K. A., and Amos, D. B. (1972)
J. Exp. Med.
136,1334-
33. Berke, J. (1983)
Immunol. Rev. 72,5-42
34. Lij~61~5%monelli, P. F., and Huan, L-J. (1988)
Methods Enzymol.
159,
35. Ohta, T., Nagano, K., and Yoshida, M. (1986)
Proc. NatL Acad. Sci. U. S.
A. 83,
2071-2075
36. Wallach D. F. H. and Ullre , D. (1962) Cancer
Res. 22, 228-234
37. Chambers D. A. ‘Salsman, 6.
Eiophyi
119,'173-178 W , and Neri, L. L. (1962)
Arch. Biochem.
38. De$$re, J. W., and Karnovsky, M. L. (1974)
J.
&cl. Ckem. 249, 7111-
39. Lin, S-H., and Russell, W. (1988)
J. BioL Chem. 263,12253-12258
40. Baner ee, R. K. (1981) Mol. Cell.
Biochem. 37.91-99
41. Dubya , G., and De Young, M. B. (1985)
J.
Brol.
Chem.
260,10653-10661 %
’ A. Filippini
and
M. V. Sitkovsky, manuscript in preparation.
by guest, on July 10, 2011www.jbc.orgDownloaded from