Tài liệu Báo cáo khoa học: Proteolysis of Pseudomonas exotoxin A within hepatic endosomes by cathepsins B and D produces fragments displaying in vitro ADP-ribosylating and apoptotic effects doc

Proteolysis of Pseudomonas exotoxin A within hepatic
endosomes by cathepsins B and D produces fragments
displaying in vitro ADP-ribosylating and apoptotic effects
Tatiana El Hage
1,2
,Se
´
verine Lorin
3
, Paulette Decottignies
4,5
, Mojgan Djavaheri-Mergny
6
and
Franc¸ois Authier
1,2
1 INSERM, Cha
ˆ
tenay-Malabry, France
2 Universite
´
Paris-Sud, Faculte
´
de Pharmacie, Cha
ˆ
tenay-Malabry, France
3 JE 2493, Universite
´
Paris-Sud, Faculte
´
de Pharmacie, Cha

ˆ
tenay-Malabry, France
4 CNRS, UMR 8619, Orsay, France
5 Universite
´
Paris-Sud, Orsay, France
6 INSERM VINCO U916, Institut Bergonie
´
, Bordeaux, France
Keywords
cathepsin; endocytosis; endosome;
Pseudomonas exotoxin A; translocation
Correspondence
F. Authier, INSERM, Universite
´
Paris-Sud,
Faculte
´
de Pharmacie, 5 rue Jean-Baptiste
Cle
´
ment, 92296 Cha
ˆ
tenay-Malabry, France
Fax: +33 1 46835844
Tel: +33 1 46835291
E-mail:
(Received 21 March 2010, revised 4 June
2010, accepted 12 July 2010)
doi:10.1111/j.1742-4658.2010.07775.x

To assess Pseudomonas exotoxin A (ETA) compartmentalization, process-
ing and cytotoxicity in vivo, we have studied the fate of internalized ETA
with the use of the in vivo rodent liver model following toxin administra-
tion, cell-free hepatic endosomes, and pure in vitro protease assays. ETA
taken up into rat liver in vivo was rapidly associated with plasma mem-
branes (5–30 min), internalized within endosomes (15–60 min), and later
translocated into the cytosolic compartment (30–90 min). Coincident with
endocytosis of intact ETA, in vivo association of the catalytic ETA-A sub-
unit and low molecular mass ETA-A fragments was observed in the
endosomal apparatus. After an in vitro proteolytic assay with an endoso-
mal lysate and pure proteases, the ETA-degrading activity was attributed
to the luminal species of endosomal acidic cathepsins B and D, with the
major cleavages generated in vitro occurring mainly within domain III of
ETA-A. Cell-free endosomes preloaded in vivo with ETA intraluminally
processed and extraluminally released intact ETA and ETA-A in vitro in a
pH-dependent and ATP-dependent manner. Rat hepatic cells underwent
in vivo intrinsic apoptosis at a late stage of ETA infection, as assessed by
the mitochondrial release of cytochrome c, caspase-9 and caspase-3 activa-
tion, and DNA fragmentation. In an in vitro assay, intact ETA induced
ADP-ribosylation of EF-2 and mitochondrial release of cytochrome c, with
the former effect being efficiently increased by a cathepsin B ⁄ cathepsin D
pretreatment. The data show a novel processing pathway for internalized
ETA, involving cathepsins B and D, resulting in the production of
ETA fragments that may participate in cytotoxicity and mitochondrial
dysfunction.
Abbreviations
DT, diphtheria toxin; EEA1, early endosome antigen-1; EF-2, elongation factor-2; ER, endoplasmic reticulum; ETA, exotoxin A;
HA, hexa-
D-arginine; LRP1, low-density lipoprotein receptor-related protein 1; PA, pepstatin-A; SD, standard deviation;
a

2
MG, a
2
-macroglobulin.
FEBS Journal 277 (2010) 3735–3749 ª 2010 The Authors Journal compilation ª 2010 FEBS 3735
Introduction
Exotoxin A (ETA) is considered to be the most toxic
factor secreted by Pseudomonas aeruginosa, a Gram-
negative opportunistic pathogen infecting immunocom-
promised individuals and burn victims [1]. ETA is a
613 amino acid A ⁄ B exotoxin that kills cells by inhibi-
tion of protein synthesis and programmed cell death
[2,3]. ETA is secreted as a single polypeptide chain
composed of three structural and functional domains:
domain Ia (amino acids 1–252), which binds to the
a
2
-macroglobulin (a
2
MG) ⁄ low-density lipoprotein
receptor-related protein 1 (LRP1) receptor present in
animal cells [4]; domain II (amino acids 253–364),
which contains a furin cleavage site (Arg276-Gln277-
Pro278-Arg279), the Cys265–Cys287 disulfide bond,
and a protein translocating sequence (amino
acids 280–313) [5,6]; and domain III (amino acids 400–
613), which contains the ADP-ribosylating enzyme [2].
To access and ADP-ribosylate its cellular target,
elongation factor-2 (EF-2), ETA must be transported
across the cellular membrane and into the cytoplasm.

This is initiated by cell surface binding of ETA to
the a
2
MG ⁄ LRP1 receptor [4], which is followed by
internalization of the toxin–receptor complex to the
endosomal apparatus by clathrin-dependent and clath-
rin-independent mechanisms [7]. Two subcellular com-
partments have been proposed as being physiologically
relevant to the mechanism of translocation of internal-
ized ETA into the cytosol. The first translocation path-
way has been proposed to operate at an early stage of
endocytosis from endocytic vesicles [8,9]. Thus, signifi-
cant translocation of ETA across the endosomal mem-
brane of mouse lymphocytes was demonstrated, and
required exposure of ETA to low endosomal pH and
ATP hydrolysis [10]. Other studies have proposed that
internalized ETA can be retrogradely transported to
the endoplasmic reticulum (ER) for retrotranslocation
to the cytosol through the Sec61 complex [11]. The ER
trafficking pathway of ETA might have multiple routes
[7], one being the previously characterized KDEL
pathway involving the REDLK C-terminal sequence
of the toxin [12].
Whatever the pathway enabling cytosolic delivery of
ETA, activating processes have been proposed to occur
at various stages of ETA trafficking. These activating
steps include furin-mediated cleavage at the Arg279-
Glu280 peptide bond [13], reduction of the disulfide
bond linking Cys265 and Cys287 [14], and removal of
the C-terminal Lys [15]. Thus, for full ADP-ribosyla-

tion of cytosolic EF-2, it was previously suggested that
intracellular production of a 37 kDa C-terminal ETA
fragment must occur by the sequential action of a
furin-like protease and an undiscovered reductive
enzyme [2,13,16]. These observations are consistent
with the toxin-resistant phenotype of cells lacking
furin, which can be abolished by transfection with a
cDNA encoding furin [17]. However, although proteo-
lytic and reductive processing of ETA should be
required for ETA cytotoxicity through the retrograde
transport pathway [18], it has not been clearly deter-
mined whether ETA requires proteolytic and ⁄ or reduc-
tive processing activation to reach the cytosol through
the endosomal pathways and kill cells [19]. Hence,
recent studies have suggested that ETA cytotoxicity
results largely from endosomal translocation of the
intact nonproteolyzed and nonreduced polypeptide
toxin [19]. At present, no in vivo data exist to support
a specificity of requirement for ETA processing and
reduction according to the translocation pathway used
(endosome or ER).
Consequently, in the present study, we used the
in situ rat liver model system following toxin adminis-
tration to rats and cell-free hepatic endosomes to relate
the endosomal processing of internalized ETA to toxin
cytotoxicity in a physiological state. Following admin-
istration of ETA to rats, rapid endocytosis of the
intact unprocessed ETA was observed, coincident with
the endosomal association of the ETA-A subunit (fast
association) and low molecular mass ETA-A fragments

(slow association). Our results assign an important role
to endosomal acidic cathepsins B and D in generating
ETA fragments displaying high in vitro ADP-ribosyl-
transferase activity towards cytosolic EF-2. We report
on the in vivo association of ETA and ETA-A with
cytosolic fractions, and the in vitro ATP-dependent
and pH-dependent translocation of ETA and ETA-A
from cell-free endosomes into the external milieu.
Finally, the mitochondrial release of cytochrome c,
activation of caspase-9 and caspase-3 and DNA frag-
mentation were detected in cytosolic fractions isolated
2 h after ETA treatment, relating for the first time
activation of the intrinsic apoptotic pathway with ETA
cytotoxicity in a physiological state.
Results
In vivo endocytosis and metabolic fate of ETA in
rat liver
The kinetics of in vivo uptake of ETA at the hepatic
cell surface (plasma membranes) (Fig. 1A) and intra-
cellularly (endosomes) (Fig. 1B) were assessed first.
Rats were given an intravenous injection of native
Proteolysis of ETA in rat hepatic endosomes T. El Hage et al.
3736 FEBS Journal 277 (2010) 3735–3749 ª 2010 The Authors Journal compilation ª 2010 FEBS
ETA (15 lg per 100 g body weight) and killed 5–90 min
postinjection. Following preparation of hepatic subcel-
lular fractions, the amount of internalized ETA was
determined by SDS ⁄ PAGE followed by western blot
analyses with antibody directed against ETA-A. It was
assumed that the in vivo generation of free ETA-A was
attributable to both reductive and proteolytic cleavages

occurring within the ETA sequence. Thus, both
processing pathways were analyzed, under either non-
reducing (cleavage analysis at the Cys265–Cys287
disulfide bond; upper blots in Fig. 1) or reducing
(cleavage analysis at peptide bonds; lower blots in
Fig. 1) conditions. ETA association with plasma mem-
branes was rapid (5 min postinjection) and maximal at
5–30 min postinjection, before decreasing with time
(Fig. 1A). A transient association of ETA-A with
plasma membranes was also observed under reducing
and nonreducing conditions at 15–60 min postinjection
(Fig. 1A). As compared with plasma membranes,
endosomal association of both ETA and ETA-A was
slightly delayed, with the maximum being observed at
15–60 min (ETA) or 30–90 min (ETA-A) (Fig. 1B).
Low molecular mass ETA fragments (< 25 kDa) were
immunodetected, especially in endosomal fractions
under reducing conditions (Fig. 1B, lower blot).
Although it has been suggested that it is the ETA–
ETA receptor complex that is internalized into toxin-
treated cells, there are no published reports on the fate
of the ETA receptor during toxin endocytosis. To
determine whether the ETA receptor was cointernal-
ized along with the toxin, the in vivo effect of ETA
treatment on the a
2
MG ⁄ LRP1 receptor in the rat liver
endosomal fraction was determined by immunoblotting
(Fig. 2A, upper blot). A high concentration of a mem-
brane-bound 80 kDa fragment of LRP1 containing the

tail epitope was found in the endosomal fraction iso-
lated from control rats. The extensive fragmentation of
LRP1 within hepatic endosomes may explain, in part,
the failure to detect intact LRP1 by us (this study) and
others [20]. In vivo injection of native ETA effected a
rapid increase in endosomal truncated LRP1, with
maximal accumulation at 5–15 min postinjection. By
60 min postinjection, the 80 kDa LRP1 species had
returned to basal levels (Fig. 2A, upper blot). How-
ever, the level of the endosomal marker early endo-
some antigen-1 (EEA1) was not modified after ETA
treatment (Fig. 2A, lower blot).
LRP1 enables endocytosis of ETA and various other
ligands among such as a
2
MG [21]. To examine the
effect of a
2
MG on the uptake of ETA into hepatic
endosomes, a
2
MG (15 lg per 100 g body weight) was
coinjected with ETA into rats (Fig. 2B). Endosomal
association of intact ETA and ETA-A was reduced by
a
2
MG coinjection.
We have previously reported that antibodies reacting
with the ER-retention KDEL motif are useful in
assessing the integrity of the C-terminal region of chol-

era toxin [22]. As it was unknown whether antibodies
Plasma membranes
Nonreducing
conditions
_
515306090
(min, postinjection)
_
515306090
Endosomes
Nonreducing
conditions
ETA
ETA
ETA-A
ETA-A
Reducing
conditions
(66 kDa) ETA
(66 kDa) ETA
Reducing
conditions
– 100
– 75
– 100
– 75
(37 kDa) ETA-A
(37 kDa) ETA-A
– 50
– 37

– 25
– 50
– 37
– 25
kDa
– 15
kDa
– 15
(min, postinjection)
AB
Fig. 1. Kinetics of appearance of ETA in hepatic plasma membranes and endosomes after toxin administration. Rat hepatic plasma mem-
brane (A) and endosomal fractions (B) were isolated at the indicated times after the in vivo administration of native ETA, and evaluated for
their content of internalized toxin by nonreducing (upper blots) and reducing SDS ⁄ PAGE (lower blots) followed by western blot analysis with
the polyclonal antibody against ETA. Each lane contained 10 lg (plasma membranes) or 30 l g (endosomes) of protein. The arrows to the left
of each panel indicate the mobilities of intact ETA ( 66 kDa), ETA-A ( 37 kDa), and unknown degradation fragments. Molecular mass
markers are indicated to the right of the reducing blots. The antibody against ETA also binds to undefined plasma membrane proteins dis-
tinct from ETA under nonreducing conditions [upper blot in (A)] in both control and toxin-injected rats, one of which had a molecular mass
identical to that of ETA-A.
T. El Hage et al. Proteolysis of ETA in rat hepatic endosomes
FEBS Journal 277 (2010) 3735–3749 ª 2010 The Authors Journal compilation ª 2010 FEBS 3737
against KDEL bind to the ETA C-terminal sequence
REDLK (which resembles the ER motif KDEL), we
first characterized antibodies against KDEL for their
binding to native ETA and ETA-A by western blot
analysis (Fig. 2C, left and middle blots). One antibody,
anti-KX
5
KDEL, bound to ETA-A but not to native
ETA (Fig. 2C, left and middle blots), whereas the
others, anti-KSEKKDEL and anti-KAVKKDEL, did

not show any immunoreactivity (results not shown).
Therefore, the antibody against KX
5
KDEL was used
to assess the integrity of the REDLK peptide in en-
dosomal ETA-A under reducing conditions (Fig. 2C,
right blot). KDEL immunoreactivity to internalized
ETA-A was detected in endosomal fractions isolated
from the livers of rats killed at 15–90 min postinjec-
tion, with kinetics similar to those of ETA-A uptake
into endocytic components (Fig. 1B), suggesting that
the C-terminal motif REDLK might not be completely
removed from ETA-A within the endosomal apparatus.
Endosomal proteolysis of internalized ETA by
cathepsins B and D.
To confirm the endosomal proteolysis of internalized
ETA under conditions that maintained endosome
integrity, we used cell-free endosomes containing
in vivo internalized ETA (Fig. 3A,B). Endosomes were
isolated 30 min following ETA injection, and intact
endocytic vesicles were incubated for various times at
ETA
_
5153060
90 (min, postinjection)
α-LRP1 (tail)
LRP1
(80 kDa)
200
*

*
100
Arbitrary units
_
5 15 30 60 90 (min, postinjection)
ETA
ETA/a
2
MG
15 30 5 15 30 (min, postinjection)
ETA-A
(37 kDa)
ETA
(66 kDa)
5
– 100
– 75
– 50
– 37
– 25
– 15
kDa
α-EEA1
EEA1
(180 kDa)
_
5 15 30 60 90 (min, postinjection)
ETA
5153060
ETA

+
_
Furin
_
ETA
(66 kDa)
+
_
Dithiothreitol
ETA-A
(37 kDa)
ETA-A
(37 kDa)
α-KX
5
KDEL
α-KX
5
KDEL
(min, postinjection)
90
A
B
C
Fig. 2. Characterization of ETA endocytosis into the endosomal
apparatus. (A) Changes in LRP1 concentration in the endosomal
fraction following ETA injection into rats. Hepatic endosomal frac-
tions were isolated at the indicated times after the in vivo adminis-
tration of native ETA, and evaluated for their content of LRP1
(upper blot) or EEA1 (lower blot) by reducing SDS ⁄ PAGE followed

by western blot analysis. Each lane contained 30 lg(a-LRP1 blot)
or 50 lg(a-EEA1 blot) of endosomal protein. The LRP1 bands were
quantified by scanning densitometry, and the signal intensities for
the ETA-treated rats were expressed as a percentage (mean ± SD)
of the signal intensity for the control rats (lane )). *P < 0.05 for the
differences between ETA ⁄ 5 min or ETA ⁄ 15 min and control rats
()). The arrows to the right indicate the mobilities of membrane-
bound LRP1 fragment ( 80 kDa) or EEA1 ( 180 kDa). Uncleaved
LRP1 ( 600 kDa) was not observed in endosomal fractions from
control and toxin-injected rats. (B) Effect of a
2
MG treatment on the
internalization of ETA. Rat hepatic endosomal fractions were iso-
lated at the indicated times after the in vivo coadministration of
ETA and a
2
MG (15 lg per 100 g body weight), and evaluated for
their content of internalized toxin by reducing SDS ⁄ PAGE followed
by western blot analysis with the polyclonal antibody against ETA.
Each lane contained 50 lg of endosomal protein. The arrows to the
left indicate the mobilities of intact ETA ( 66 kDa), ETA-A
( 37 kDa), and unknown degradation fragments. Molecular mass
markers are indicated to the right. (C) Assessment of immunoreac-
tivity of antibody against KDEL for native and internalized ETA. ETA
was either untreated (left blot, lane ) ) or digested in vitro with
100 UÆmL
)1
Æmg
)1
furin and 10 mM dithiothreitol (middle blot,

lane +), and samples were then analyzed by reducing SDS ⁄ PAGE
followed by western blotting with polyclonal antiserum against the
synthetic peptide KX
5
KDEL. ETA-A was detected under the latter
experimental conditions. Rat liver endosomal fractions were then
isolated at the indicated times after the in vivo administration of
ETA, and evaluated by western blotting for their immunoreactivity
with polyclonal antibody against KX
5
KDEL (blot on the right) [22].
The antibody against KX
5
KDEL also binds to undefined endosomal
proteins distinct from ETA, both in control and in toxin-injected rats,
whose levels have been shown to be modified by toxin treatment
[22]. Each lane contained 80 lg of endosomal protein. The mobili-
ties of intact ETA ( 66 kDa) and ETA-A ( 37 kDa) are indicated.
Proteolysis of ETA in rat hepatic endosomes T. El Hage et al.
3738 FEBS Journal 277 (2010) 3735–3749 ª 2010 The Authors Journal compilation ª 2010 FEBS
neutral pH (pH 7) and 37 °C in an isotonic buffer
(which mimicked the intracellular milieu) in the pres-
ence or absence of ATP, the substrate of the vacuolar
H
+
-ATPase pump (Fig. 3A). Immunoblot analyses
showed a progressive loss of intact ETA and ETA-A
in the presence of ATP, with concomitant generation
of ETA and ETA-A fragments. Incubation in the
absence of ATP revealed a small amount of degrada-

tion for intact ETA, whereas no degradation was
observed for ETA-A (Fig. 3A).
We next examined the effects of various protease
inhibitors on the proteolysis of endosomal ETA and
ETA-A, using cell-free endosomes preloaded with ETA
toxin in vivo and incubated in vitro at pH 7 in the pres-
ence of ATP (Fig. 3B). Western blot analysis with the
antibody against ETA revealed that the endosomal
ETA-degrading activity was partially inhibited by the
aspartic acid protease inhibitor pepstatin-A (PA), the
cysteine protease inhibitor E64, and the metallopro-
tease inhibitor EDTA.
The inhibition of ETA-degrading activity by PA and
E64, its low pH optimum and its presence in the
endosomal lumen as a soluble form (results not shown)
suggested cathepsins B and D as likely candidates for
this activity. We therefore examined the hydrolysis of
ETA by pure cathepsins B and D at pH 4–7 (Fig. 3C).
pH 7 + ATP
_
PA
HA
EDTA
PMSF
E64
_
06060606060
60 (min)
(Inhibitor)
(

37 kDa) ETA-A
(66 kDa) ETA
ETA + Cath-D ETA + Cath-B
44554455667
7pH
15 60 15 60
15
60 15 60 15 60 15
60 (min)
_
ETA
(66 kDa)
ETA-A
(37 kDa
)
0 15 30 0 15 30 (min)
pH 7 pH 7 + ATP
ETA-A (37 kDa)
ETA (66 kDa)
100 –
75 –
50 –
37 –
25 –
kDa
15 –
150
ETA ETA-A subunit
0 0 30 30 30 30 (min)
50

100
Arbitrary units
_
+
_
+
_
+ (ATP)
A
B
C
Fig. 3. Assessment of ETA-degrading activity associated with
hepatic endosomes. (A) Rat hepatic endosomal fractions were iso-
lated 30 min after ETA administration (15 lg per 100 g body
weight) and incubated for the indicated times at 37 °C in isotonic
buffer containing 0.15
M KCl, 25 mM Hepes (pH 7), 5 mM MgCl
2
,
and 6 m
M CaCl
2
, in the presence or absence of 10 mM ATP. The
integrity of ETA was then evaluated by reducing SDS ⁄ PAGE fol-
lowed by western blotting with the polyclonal antibody against
ETA. Each lane contained 50 lg of endosomal protein. The arrows
to the right indicate the mobilities of intact ETA ( 66 kDa), ETA-A
( 37 kDa), and unknown degradation fragments. Molecular mass
markers are indicated to the left. ETA and ETA-A signals were
quantified by scanning densitometry, and the signal intensities after

a 30 min incubation were expressed as a percentage (mean ± SD)
of initial values (0 min) (lower panel). (B) Rat hepatic endosomal
fractions were isolated 30 min after ETA administration (15 lg per
100 g body weight) and incubated at 37 °C in isotonic buffer con-
taining 0.15
M KCl, 25 mM Hepes (pH 7), 5 mM MgCl
2
,6mM CaCl
2
and 10 mM ATP for the indicated times in the presence or absence
(lane ))of2m
M phenylmethanesulfonyl fluoride (PMSF),
10 lgÆmL
)1
PA, 5 mM EDTA, 10 lM HA, or 10 lM E64. The integrity
of ETA was then evaluated by reducing SDS ⁄ PAGE followed by
western blotting with the polyclonal antibody against ETA. Each
lane contained 50 lg of endosomal protein. The arrows to the left
and right indicate the mobilities of intact ETA ( 66 kDa), ETA-A
( 37 kDa), and unknown degradation fragments. (C) Native ETA
(10 lg) was incubated at 37 °C with cathepsin D (Cath-D) or
cathepsin B (Cath-B) (5 UÆmL
)1
Æmg
)1
)in50mM citrate ⁄ phosphate
buffer (pH 4–6) or 50 m
M Hepes (pH 7) in the presence of 10 mM
CaCl
2

and 10 mM dithiothreitol for the indicated times. The integrity
of ETA was then evaluated by reducing SDS ⁄ PAGE followed
by western blotting with the polyclonal antibody against ETA.
The arrows to the left and right indicate the mobilities of intact
ETA ( 66 kDa), ETA-A ( 37 kDa), and unknown degradation
fragments.
T. El Hage et al. Proteolysis of ETA in rat hepatic endosomes
FEBS Journal 277 (2010) 3735–3749 ª 2010 The Authors Journal compilation ª 2010 FEBS 3739
Western blot analysis with an antibody against ETA
showed that cathepsins B and D degraded ETA in a
pH-dependent manner, with maximal degradation
being observed at pH 4. The ETA fragments generated
by the pure cathepsins (especially cathepsin B at pH 4)
had molecular masses very similar to those seen with
the endosomal fractions.
We then assessed the major proteolytic cleavages
induced by cathepsin B and ⁄ or D within the ETA
sequence at various pH values (Fig. 4A,B). The prote-
olysis of ETA at pH 4 or 6 by cathepsin B and ⁄ or D
was analyzed by reducing SDS⁄ PAGE (Fig. 4A), and
the cleavage sites in the major metabolites were deter-
mined by N-terminal sequence analysis (Fig. 4B).
Edman degradation of intermediates 4a, 4b, 4d, 6a
and 6c revealed the N-terminal sequence of ETA
(AEEAFDL), suggesting that the cleavage sites are
located within the C-terminal region of the toxin.
N-terminal sequence analysis of ETA fragments 6b
and 6d, generated at pH 6, revealed cleavages
between Thr396 and Cys397 (as demonstrated by the
CPVAAGECA sequence). For peptide 4c, generated

at pH 4, N-terminal sequence analysis revealed the
peptide PALA, suggesting cleavage between Asp499
and Pro500.
Assessment of cytosolic translocation of
internalized ETA
We next determined the presence of ETA in cytosolic
fractions prepared from ETA-injected rats, using wes-
tern blot analysis (Fig. 5A). The intact 66 kDa ETA
toxin was strongly detected within cytosolic fractions
at 0.5–4 h postinjection, and lower but detectable
immunoreactivity for ETA-A was observed at 1–4 h
under both reducing and nonreducing conditions. The
translocation of endosomal ETA into the extraendoso-
mal milieu was then assessed with intact endosomes
isolated 30 min after the injection of ETA and incu-
bated for 0–2 h in isotonic medium at 37 °C in the
presence or absence of ATP (Fig. 5B). Western blot
analysis of the ETA associated with sedimentable
endosomes showed progressive decreases in immunore-
active ETA and ETA-A at acidic pH (pH 5) or at
pH 7 in the presence of ATP. Concomitantly, immu-
noreactive ETA (high level) and ETA-A (low level)
were progressively detected in the extraendosomal
milieu, confirming the translocation of ETA toxin
across the endosomal membrane at acidic luminal pH.
No ATP-dependent translocation of ETA was observed
in the presence of b afilomycin, the H
+
-ATPase inhibitor.
– 100

– 75
kDa
ETA
(66 kDa)
+ + + + Cathepsin B
+ + + Cathepsin D
4 6 4 6 4 pH of medium
ETA
1
– 50
– 37
– 25
– 15
4b
4c
6c
4a
6b
6a
6d
4d
0.25 3 0.25 3 0.25
Incubation time (h)
AEEAFDLWNECAKACVLDLKDGVRSSRMSVDPAIADTNGQGVLHYSMVLEGGNDALKLAIDN
ALSITSDGLTIRLEGGVEPNKPVRYSYTRQARGSWSLNWLVPIGHEKPSNIKVFIHELNAGN
QLSHMSPIYTIEMGDELLAKLARDATFFVRAHESNEMQPTLAISHAGVSVVMAQTQPRREKR
WSEWASGKVLCLLDPLDGVYNYLAQQRCNLDDTWEGKIYRVLAGNPAKHDLDIKPTVISHRL
HFPEGGSLAALTAHQACHLPLETFTRHRQPR
279
1

ETA-B
280
GWEQLEQCGYPVQRLVALYLAARLSWNQVDQ
V
IRNALASPGSGGDLGEAIREQPEQARLALTLAAAESERFVRQGTGNDEAGAANADVVSLTCP
VAAGECAGPADSGDALLERNYPTGAEFLGDGGDVSFSTRGTQNWTVERLLQAHRQLEERGYV
FVGYHGTFLEAAQSIVFGGVRARSQDLDAIWRGFYIAGDPALAYGYAQDQEPDARGRIRNGA
LLRVYVPRSSLPGFYRTSLTLAAPEAAGEVERLIGHPLPLRLDAITGPEEEGGRLETILGWP
ETA-A
LAERTVVIPSAIPTDPRNVGGDLDPSSIPDKEQAISALPDYASQPGKPPREDLK
613
A
B
Fig. 4. Structural characteristics of ETA fragments generated by cathepsin B and cathepsin D. (A) Native ETA (10 lg) was incubated with
bovine cathepsin B or cathepsin D (5 UÆmL
)1
Æmg
)1
), or a mixture of both, at 37 °C for the indicated times in 50 mM citrate ⁄ phosphate buffer
(pH 4–6). The incubation mixtures were then analyzed by reducing SDS ⁄ PAGE followed by Coomassie Brilliant Blue Staining. The major deg-
radation products generated at pH 4 (peptides 4a–c) or pH 6 (peptides 6a–d) were subjected to N-terminal sequence analysis. (B) Sequences
of ETA-A and ETA-B. The A and B moieties are connected by both a peptide bond (Arg279-Gly280) and a disulfide bridge (Cys265–Cys287).
The peptides in red correspond to the N-terminal sequence of intermediates shown in (A): AEEAFDL for intermediates 1, 4a, 4b, 4d, 6a and
6c; CPVAAGECA for intermediates 6b and 6d; and PALA for intermediate 4c.
Proteolysis of ETA in rat hepatic endosomes T. El Hage et al.
3740 FEBS Journal 277 (2010) 3735–3749 ª 2010 The Authors Journal compilation ª 2010 FEBS
Potential role of cathepsins B and D in the
cytotoxic activity of ETA towards cytosolic targets
We first examined whether, under conditions where
ETA was processed by cathepsins B and D, a corre-

sponding change in the toxin cytotoxicity towards
cytosolic EF-2 would be observed (Fig. 6A). ETA was
first partially processed by a mixture of cathepsins B
and D at pH 4 or 6, and then incubated at neutral pH
with cytosolic EF-2 in the presence of [
32
P]NAD. A
low level of ADP-ribosylation of EF-2 was evident
after addition of untreated ETA to the cytosolic frac-
tion. After treatment of ETA with cathepsins B and
D, EF-2 labeling was increased, especially under acidic
conditions (pH 6 > pH 5 > pH 4). However, cathep-
sin treatment of ETA in the presence of protease
inhibitors revealed [
32
P]NAD-ribose incorporation into
cytosolic EF-2 comparable to that observed in the
absence of protease treatment.
A role for mitochondria in ETA-induced cell death
has been previously shown with the use of human air-
way epithelial target cells [23]. Consequently, we exam-
ined cytochrome c release from cell-free mitochondria
isolated from control rats and then treated with ETA
in vitro (Fig. 6B, upper blots). Cytochrome c associa-
tion with intact rat liver mitochondria persisted during
the incubation in isotonic medium, despite small but
detectable release at 15 min. However, there was sub-
stantial release of cytochrome c into the resulting mito-
chondrial supernatant after the addition of native ETA
or ETA that had been pretreated with a mixture of

cathepsins B and D. No detectable release of cyto-
chrome c was observed following treatment of mito-
chondria with a mixture of cathepsins B and D alone
for the same incubation times (results not shown).
To assess the physiological release of mitochondrial
cytochrome c into the cytosol, hepatic cytosolic frac-
tions isolated after the in vivo injection of ETA into rats
were analyzed for their cytochrome c content by immu-
noblot analysis (Fig. 6B, lower blots). Low but detect-
able immunoreactivity towards cytochrome c was
observed in cytosol isolated from noninjected rats. In
response to ETA, a strong increase in cytochrome c was
observed at 2 h, with the level remaining elevated up to
4 h. By contrast, administration of diphtheria toxin
(DT) (a toxin that does not access the cytoplasm of
rodent cells [24]) did not cause a detectable change in the
level of cytochrome c in the cytoplasmic compartment.
The involvement of caspases in ETA-triggered pro-
grammed cell death was then analyzed by incubating
hepatic cytosol isolated from ETA-treated rats with
fluorescent substrates specific for caspase-9, caspase-3
and caspase-8 (Fig. 6C, open bars). Caspase-9 and
_
0.5 1 2
Pellet
(endosomal medium)
_
0.5 1 2 (incubation, h)
Supernatant
(extraendosomal medium)

Cytosol
ETA-A
(37 kDa)
(66 kDa) ETA
Medium: pH 7
_
0.5 1 1.5 2 4 (h, postinjection)
ETA
Nonreducing
conditions
ETA-A
ETA-A
(37 kDa)
(66 kDa) ETA
Medium: pH 7 + ATP
ETA-A
(37 kDa)
ETA
(66 kDa)
Reducing
conditions
ETA-A
(37 kDa)
(66 kDa) ETA
Medium: pH 5
ETA-A
(37 kDa)
(66 kDa) ETA
Medium: pH 7 + ATP + Bafilomycin
A

B
Fig. 5. In vivo and in vitro assessment of the cytosolic transloca-
tion of endosomal ETA. (A) Rat hepatic cytosolic fractions were iso-
lated at the indicated times after the in vivo administration of native
ETA, and evaluated for their toxin content by nonreducing (upper
blot) and reducing (lower blot) SDS ⁄ PAGE followed by western blot
analysis with the polyclonal antibody against ETA. Each lane con-
tained 30 lg of cytosolic protein. The arrows to the left indicate the
mobilities of intact ETA ( 66 kDa) and ETA-A ( 37 kDa). (B)
Membrane translocation of toxin peptides in cell-free rat hepatic
endosomes containing in vivo internalized ETA. The endosomal
fraction was isolated 30 min after the administration of ETA, and
then resuspended in 0.15
M KCl containing 5 mM MgCl
2
and, when
indicated, 50 m
M Hepes (pH 7), 50 mM citrate ⁄ phosphate buffer
(pH 5), 10 m
M ATP, and 1 lM bafilomycin. After the indicated times
of incubation at 37 °C, endosomes were sedimented by ultracen-
trifugation, and the pellets (endosome-associated material) and
supernatants (extraendosomal material) were evaluated for their
content of ETA peptides by reducing SDS ⁄ PAGE followed by wes-
tern blotting with the polyclonal antibody against ETA. Equivalent
volumes of each subfraction (40 lL) were loaded onto each lane.
The arrows to the left indicate the mobilities of intact ETA
( 66 kDa) and ETA-A ( 37 kDa).
T. El Hage et al. Proteolysis of ETA in rat hepatic endosomes
FEBS Journal 277 (2010) 3735–3749 ª 2010 The Authors Journal compilation ª 2010 FEBS 3741

no toxin
Native
ETA
Cathepsin-treated
ETA
EF-2
(105 kDa)
_
7654
Cathepsin-treated ETA + cytosol
4+i (pH of proteolysis)
ADP-ribosylation of EF-2
Cytochrome c
(15 kDa)
Cytochrome
c
(15 kDa)
_
515
_
515
_
5 15 (min of incubation)
Intact
mitochondria
Disrupted
mitochondria
EF-2
(105 kDa)
WB: α-EF-2

100
200
*
*
_
0.5 1 1.5 2 4 (h, postinjection)
Cytochrome c
(15 kDa)
ETA
Caspase-9
2
ETA DT
Toxin
Arbitrary units
(ratio of [
32
P]EF2/EF2)
_
7 6 5 4 4+i (pH of proteolysis)
DT
Cytochrome c
(15 kDa)
1
Caspase-3
4
ETA DT
2
4
Caspase activity (fold stimulation)
Caspase-8

2
1
Cell death (fold stimulation)
_
5 15 30 60 90 120 240(min, postinjection)
_
0.5 1 1.5 2 4 (h, postinjection)
0.5
A
C
D
B
Fig. 6. Assessment of cytotoxic activity of cathepsin-treated ETA towards cytosolic target and mitochondria. (A) Native ETA (10 lg) was
digested in vitro for 30 min at 37 °C with a mixture of cathepsins B and D (5 UÆmL
)1
Æmg
)1
)in25mM Hepes (pH 7) or 25 mM citrate ⁄ phos-
phate buffer (pH 4–6) containing 0.1
M dithiothreitol (DT) and, when indicated, 5 lgÆmL
)1
PA and 1 lM E64 (lane 4 + i). The treated ETA
(1 lg) was then incubated for 15 min at 30 °C with the EF-2 associated with the soluble cytosolic fraction (150 lg) in 0.1
M Hepes (pH 7.4)
in the presence of 2 l
M [
32
P]NAD. Samples (20 lg) were then subjected to SDS ⁄ PAGE and analyzed by autoradiography. The dried gels
were exposed to X-ray film at )80 °C for 1–3 days. The arrow to the left indicates the mobility of
32

P-labeled EF-2 ( 105 kDa). Samples
(20 lg) were also evaluated for their content of EF-2 using polyclonal antibody against EF-2. The arrow to the left indicates the mobility of
EF-2 ( 105 kDa). For each incubation condition, radiolabeled and nonradiolabeled EF-2 signal intensities were quantified by scanning densi-
tometry, and the ratio of [
32
P]EF-2 signal ⁄ nonradiolabeled EF-2 signal was expressed as a percentage (mean ± SD) of that of untreated ETA
(lane ), 100%) (lower panel). *P < 0.05 for the differences between pH 5 or pH 4 and untreated cytosol. (B) Upper blots: rat liver mitochon-
dria (7.5 mgÆmL
)1
) were incubated in either 0.15 M KCl isotonic buffer (intact mitochondria, blot at the top) or hypotonic buffer (disrupted
mitochondria, lower blot) in the presence or absence of native or cathepsin-treated ETA. After the indicated times, samples were centrifuged
and mitochondrial supernatants were carefully separated and mixed with sample buffer. Equivalent volumes (30 lL) were subjected to
reducing SDS ⁄ PAGE followed by western blot analysis for the in vitro release of cytochrome c. The arrows to the right indicate the mobility
of cytochrome c ( 15 kDa). Lower blots: rat hepatic cytosolic fractions were isolated at the indicated times after the in vivo administration
of native ETA or diphtheria toxin (DT), and evaluated by western blotting with monoclonal antibody for their content of cytochrome c. Each
lane contained 30 lg of cytosolic protein. The arrows to the right indicate the mobility of cytosolic cytochrome c ( 15 kDa). (C) Hepatic
cytosolic fractions isolated from ETA-injected or DT-injected rats were incubated with fluorescent substrates specific for caspase-9,
caspase-3, and caspase-8. The results are expressed as fold stimulation of fluorescence intensity, normalized to that seen in the control rats,
and represent the mean ± SD of three determinations performed on the cytosolic fraction prepared from separate liver fractionations.
(D) Histone-associated DNA fragments associated with hepatic cytosolic fractions isolated from ETA-injected and DT-injected rats were
analyzed by immunoassay. Results are expressed as fold stimulation, normalized to that seen in the control rats, and represent the mean of
two determinations performed on the cytosolic fractions prepared from separate liver fractionations.
Proteolysis of ETA in rat hepatic endosomes T. El Hage et al.
3742 FEBS Journal 277 (2010) 3735–3749 ª 2010 The Authors Journal compilation ª 2010 FEBS
caspase-3 activity increased in rat liver cytosol 1.5–2 h
after the injection of ETA, with a maximal effect of
 2.7-fold (caspase-9) or 4.0-fold (caspase-3) at 4 h.
No activation of caspase-8 (involved in the extrinsic
death receptor pathway) was observed. No increase in
caspase activity was observed in hepatic cytosolic frac-

tions isolated from DT-injected rats (Fig. 6C, closed
bars). Finally, the kinetics and extent of production of
histone-associated DNA fragments in hepatic cytosolic
fractions following ETA administration into rats paral-
leled caspase-9 and caspase-3 activation, with DNA
fragmentation being observed 2 h after ETA injection
and remaining elevated up to 4 h (Fig. 6D, open bars).
No DNA fragmentation was observed in hepatic cyto-
solic fractions isolated from DT-injected rats (Fig. 6D,
closed bars).
Discussion
Using the in situ liver model system, we have previ-
ously shown that, after cholera toxin binding to hepa-
tic cells, cholera toxin accumulates in a low-density
endosomal compartment and then undergoes endoso-
mal proteolysis by the aspartic acid protease cathep-
sin D [22,25]. Using a similar methodology, others
have previously shown that the plant toxin ricin fol-
lows a similar intraendosomal processing pathway,
requiring ATP-dependent endosomal acidification [26].
We have recently extended these observations to DT,
and demonstrated the endosomal processing of the
internalized toxin in a sequential degradation pathway
beginning early, prior to organelle acidification via a
neutral furin activity, and followed later under acidic
conditions via cathepsin D [24]. In the present work,
we have evaluated the relationship between the
endosomal processes and cytotoxicity of ETA, another
A ⁄ B toxin functionally related to DT that has an iden-
tical intracellular target (cytosolic EF-2) [6]. Our data

clearly show that internalized ETA is susceptible to
hydrolysis by cathepsins B and D, which are present in
hepatic endosomes and operate at acidic pH. Compa-
rable to the endosomal degradation of internalized CT
[22,25] and ricin [26] in rat hepatic endosomes, the
endosomal processing of internalized ETA occurred
mainly (if not totally) following ATP-dependent acidi-
fication of the endosomal lumen.
Cytosolic translocation of endosomal ETA was
established through the immunodetection of the toxin
in cytosol isolated from ETA-injected rats and the use
of cell-free endosomes. Thus, intact ETA and ETA-A
were the only ETA species detected in vivo in the
soluble cytosolic fraction after toxin administration
and in vitro in the extraendosomal medium during a
cell-free translocation assay. However, we cannot
exclude the possibility that a small number of ETA
fragments generated by endosomal cathepsins B and D
physiologically translocate from the endosomal lumen
to the cytoplasm and interact with cytosolic targets
(EF-2 and mitochondria). Low production and ⁄ or
translocation of ETA fragments, as well as short half-
lives in the cytosolic compartment, may well explain
why they were not detected. Alternatively, the pro-
cessed fragments may have lost structural elements
essential for translocation across the endosomal mem-
brane. On the other hand, endosomal proteolysis of
ETA may represent a degradative pathway related to
the deactivation and termination of intracellular toxin
cytotoxicity. Clearly, further studies are required to

determine whether ETA fragments generated by en-
dosomal cathepsins B and D fully participate in the
cytotoxic action of ETA in hepatic tissue.
Intravenously injected ETA is taken up efficiently by
the liver at an early time after death (5 min postinjec-
tion), suggesting a high binding capacity of ETA in
hepatic parenchyma. Indeed, injection of the toxin into
mice has been shown to result in an early and pro-
found inhibition of hepatic protein synthesis [27]. Our
results suggest that a
2
MG ⁄ LRP1 contributes, at least
in part, to ETA endocytosis in rat liver in vivo, based
on the following: (a) the injection of a
2
MG, which par-
tially reduced the endosomal association and process-
ing of coinjected ETA; and (b) a time-dependent
increase in immunodetectable a
2
MG ⁄ LRP1 in hepatic
endosomes induced by the toxin injection.
It has been proposed that proteolytic nicking of
ETA at the Arg279-Glu280 peptide bond mediated by
furin activity is at least partly required for expression
of ETA cytotoxicity [2,13]. In the present study, our
observation that ETA-A associates with hepatic
plasma membrane, endosomal and cytosolic fractions
isolated from ETA-injected rats is consistent with this
view. However, our in vivo and in vitro data also sup-

port the contention that the furin-mediated conversion
of native ETA into ETA-A within hepatic endosomes
may represent a minor metabolic fate for the internal-
ized toxin, based on the following: (a) the lack of
sensitivity of endosomal ETA-degrading activity to
furin inhibitors [hexa-d-arginine (HA)]; and (b) the
predominant association of low molecular mass frag-
ments of ETA-A with hepatic endosomes at a late
stage of ETA endocytosis (60 min post-ETA treat-
ment). Finally, our data suggesting the presence of
intact ETA and ETA-A at the cell surface are consis-
tent with the endocytosis of native ETA (major
pathway) as well as ETA-A (minor pathway) from the
cell surface to early endosomes [28,29].
T. El Hage et al. Proteolysis of ETA in rat hepatic endosomes
FEBS Journal 277 (2010) 3735–3749 ª 2010 The Authors Journal compilation ª 2010 FEBS 3743
In the present work, we showed that the endosomal
acidic proteolytic activity directed towards the internal-
ized ETA was comparable to that of the cysteine pro-
tease cathepsin B and the aspartic acid protease
cathepsin D, as indicated by the following observa-
tions: (a) the inhibitor profile of the endosomal
ETA-degrading activity was very similar to that of
cathepsins B and D [30]; and (b) the endosomal activ-
ity produced a substrate cleavage pattern that was very
similar to that generated with pure cathepsins B or D.
Interestingly, previous studies have shown that intra-
cellular processing of ETA by a PA-sensitive protease
was critical for ETA-induced lymphoproliferation,
confirming that one or more intracellular proteases

distinct from furin participate in ETA processing
within toxin-treated cells [31]. Moreover, additional
metallo-dependent proteolytic activities (EDTA-sensi-
tive) might act on internalized ETA within endosomes
and produce fragments with a molecular mass very
close to that of intact ETA.
All cleavages produced by cathepsins B and D in the
ETA toxin are located within ETA-A. A major degra-
dation product of ETA results from proteolytic cleav-
age at Thr396-Cys397 in the C-terminal extremity of
domain I or Ib. The degradation product contains the
entire catalytic ETA-A domain (amino acids 400–613)
extended at the N-terminus by the CPV tripeptide, and
may represent the main catalytic fragment respon-
sible for the ADP-ribosyltransferase activity identified
in vitro after cathepsin treatment. Three degradation
products (peptides 4a, 4b and 4d) displayed a molecu-
lar mass slightly less than that of the native 66 kDa
ETA and the unmodified N-terminal ETA sequence,
suggesting the removal of the C-terminal residues of
ETA encompassing the REDLK sequence. However,
an antibody that recognizes the REDLK-mediated ER
retrieval motif, which is located at the C-terminus of
ETA-A, showed immunoreactivity with endosomal
ETA-A, suggesting that the REDLK motif was not
completely lost from ETA-A within endosomes. It has
previously been shown that human serum contains a
carboxypeptidase activity, suggested to be carboxypep-
tidase-N, carboxypeptidase-H or carboxypeptidase-M,
which removed the C-terminal Lys of ETA and gener-

ated a processed form of ETA ending in REDL [15].
We have now extended these observations to the
endosomal apparatus, and suggest that ETA may
undergo C-terminal processing that begins early in the
circulating blood, and is continued later within endo-
somes after entry of the toxin into the cell.
Western blot analyses of ETA associated with hepa-
tic subcellular fractions under nonreducing conditions
showed that the Cys265–Cys287 disulfide bridge was
partially cleaved at the plasma membrane, endosome
and cytosol loci. Thus, as for the proteolytic cleavage
of ETA at the connecting A ⁄ B junction bond, the
hepatic ETA-reducing activity may well operate early
at the cell surface prior to ETA endocytosis. More-
over, the level of ETA reduction within hepatic endo-
somes was much lower than that of proteolysis,
suggesting that the endosomal reductive pathway may
represent a minor metabolic fate for the internalized
toxin [32]. It has been previously suggested that ETA
reduction is a two-step process: toxin unfolding that
allows access to the Cys265–Cys287 bond is followed
by reductive cleavage of the disulfide bond by a pro-
tein disulfide isomerase-like enzyme [14]. Importantly,
toxin unfolding and reducing activities were present in
the membrane fraction of toxin-sensitive cells but not
in the soluble fraction, suggesting that the cytosol and
the endosomal lumen may not be the relevant com-
partments for such cell-mediated reducing events [14].
One endosome-located mechanism that regulates
ETA activation and action occurs at the level of orga-

nelle acidification [33]. First, a low pH has been pro-
posed to be required for the proteolytic cleavage of
ETA by furin [34]. Thus, whereas furin displays an
optimal pH of  7 for model peptide substrates [35],
the proteolysis of ETA by furin is maximal between
pH 5.0 and pH 5.5 [34]. Moreover, the vacuolar
H
+
-ATPase inhibitor bafilomycin protected mouse
L cells from the toxic effects of intact ETA as well as
precleaved ETA, suggesting that an acidic environment
is required for proteolytic activation of ETA and addi-
tional event(s) leading to its cytotoxic effect [33].
Finally, it has clearly been shown that endosomal acid-
ity facilitates the binding of ETA to the endosomal
membrane of mouse L cells (maximal binding observed
at pH 4.0) and ETA-induced pore formation in the
lipid bilayer of endosomal vesicles (maximal effect at
pH < 6) [8]. Our data showing the in vitro proteolysis
of ETA by endosomal acidic cathepsins and transloca-
tion of the internalized toxin across the endosomal
membrane at low pH would be consistent with these
prior observations. Other studies reported that ETA
translocation was strictly dependent on ATP hydrolysis
but was not affected by bafilomycin, the H
+
-ATPase
inhibitor [9]. These differences may result from the
experimental approaches used (the rat liver in vivo
model versus cellular in vitro systems) and ⁄ or may be

related to differences between hepatocytes and other
cell types.
In vivo [36] and in vitro [37] studies have shown that
the normal airway epithelium is highly resistant to
P. aeruginosa-induced apoptosis. Moreover, in airway
target cells, ETA induced a wide range of biochemical
Proteolysis of ETA in rat hepatic endosomes T. El Hage et al.
3744 FEBS Journal 277 (2010) 3735–3749 ª 2010 The Authors Journal compilation ª 2010 FEBS
and morphological changes (early mitochondrial dys-
function) that are not characteristic of apoptosis [23].
On the other hand, live P. aeruginosa bacteria can
induce the apoptotic death of human conjunctiva
epithelial Chang cells [38] and J774A.1 macrophages
[39] through the type III secretion system. Also,
ETA-induced human mast cell apoptosis by activation
of the caspase-8 and caspase-3 pathways and downre-
gulation of antiapoptotic FLIP proteins has been
reported [3]. In the present work, we have demon-
strated that, after ETA administration, rat hepatic
cells undergo in vivo apoptosis through DNA frag-
mentation, mitochondrial release of cytochrome c, and
activation of caspase-9 and caspase-3. By contrast,
the receptor caspase-8-dependent pathway did not
contribute to ETA-induced apoptosis in rat liver cells
in vivo.
We speculate that, in the cytoplasm of toxin-trea-
ted hepatic cells, translocated ETA can effect its
cytotoxicity through a dual action, i.e. ADP-ribosyla-
tion of EF-2 (inhibition of protein synthesis) and
mitochondrial alteration (intrinsic apoptotic effect).

Both pathways require the translocation of ETA into
the cytoplasm of toxin-treated cells. On the basis of
the reconstitution of the cytotoxic pathways with
in vitro cytosol and cell-free mitochondria, our data
suggest a direct interaction between ETA and cyto-
solic EF-2 on the one hand, and the mitochondrial
membrane on the other hand. However, the potential
role (if any) of ADP-ribosylation of EF-2 in the
mitochondrial apoptotic response induced by the
toxin remains to be determined. Finally, we assign
an important role to the endosomal acidic cathep-
sins B and D in increasing the in vitro transfer of
the ADP-ribosyl moiety of NAD
+
to EF-2 by ETA,
but not in the release of cytochrome c from cell-free
mitochondria.
In summary, we found that internalized ETA was
rapidly proteolyzed within rat hepatic endosomes by
cathepsins B and D, with subsequent ATP-dependent
translocation of intact ETA and ETA-A to the cytosol.
Intact ETA induced ADP-ribosylation of cytosolic
EF-2 as well as the mitochondrial release of cyto-
chrome c, both in vivo and in vitro, with the in vitro
effects being substantially increased by cathepsin B ⁄ D
pretreatment. Studies are currently underway to eluci-
date whether ETA-induced mitochondrial alteration is
mediated by the catalytic A-subunit or hydrophobic
B-domain of ETA, or whether it involves the dual het-
erogeneous part of the toxin. Use of this approach will

provide novel insight(s) into the physiological signifi-
cance of the endosomal fragments of internalized
ETA, which, until now, has remained elusive.
Experimental procedures
Peptides, enzymes, antibodies, protein determi-
nation, N-terminal sequencing, enzyme assays,
and materials
Pseudomonas aeruginosa ETA, DT and bafilomycin-A1
were purchased from Calbiochem. Bovine cathepsin D
(EC 3.4.23.5, 15 UÆmg
)1
), bovine cathepsin B (EC 3.4.22.1,
10 UÆmg
)1
), recombinant truncated human furin
(2000 UÆmL
)1
) and human plasma a
2
MG were purchased
from Sigma. Rabbit antibody against Pseudomonas ETA
was purchased from Sigma. Western blot analysis using the
antibody against ETA revealed a strong affinity for ETA
with a specificity for the A-subunit. Mouse monoclonal
antibody directed against rat EEA1 was purchased from
Transduction Laboratories. Mouse monoclonal antibody
directed against rat cytochrome c was purchased from
Pharmingen. Rabbit polyclonal antibody against human
EF-2 and goat polyclonal antibody raised against the
C-terminus of human LRP1 were purchased from Santa

Cruz Biotechnology. Rabbit polyclonal antibody against
KX
5
KDEL, which recognizes the ER retention signal
KDEL and binds to various ER-resident proteins, was
obtained from S. Fuller (EMBL, Heidelberg, Germany).
Horseradish peroxidase-conjugated goat anti-(rabbit IgG)
or goat anti-(mouse IgG) were from Sigma. The protein
content of isolated fractions was determined by the method
of Lowry et al. [40]. N-terminal sequence data were
obtained by automated Edman degradation with a Procise
sequencer (Applied Biosystems, Foster City, CA, USA),
equipped with an on-line phenylthiohydantoin amino acid
analysis system. Quantitative analysis of DNA fragmenta-
tion after toxin-induced cell death was analyzed by immu-
noassay determination of cytoplasmic histone-associated
DNA fragments, according to the manufacturer’s protocol
(Roche). N-Acetyl-b-d-glucosaminidase was assayed with
p-nitrophenyl N-acetyl-b-d-glucosaminide as substrate,
according to Touster et al. [41]. Acid phosphatase was
assayed as described by Trouet [42]. Caspase-3, caspase-8
and caspase-9 activity was analyzed with a fluorometric
assay kit (BioVision, Mountain View, CA, USA) with the
respective DEVD-AFC, IETD-AFC and LEHD-AFC sub-
strates. Nitrocellulose membranes and the enhanced chemi-
luminescence detection kit were from Amersham. PA, E-64,
phenylmethanesulfonyl fluoride and EDTA were from
Sigma. HA was from Calbiochem. All other chemicals were
obtained from commercial sources and were of reagent grade.
Animals and injections

In vivo procedures were approved by the institutional
committee for use and care of experimental animals.
Male Sprague-Dawley rats, body weight 180–200 g, were
obtained from Charles River France (St Aubin Les Elbeufs,
T. El Hage et al. Proteolysis of ETA in rat hepatic endosomes
FEBS Journal 277 (2010) 3735–3749 ª 2010 The Authors Journal compilation ª 2010 FEBS 3745
France) and fed ad libitum. Native ETA or DT (15 lg per
100 g body weight) in 0.4 mL of 0.15 m NaCl was injected
within 5 s into the penile vein under light anesthesia with
ether.
Rat liver subcellular fractionation
Subcellular fractionation was performed using established
procedures [22,24,25]. Following injection of toxins, rats
were killed and livers were rapidly removed and minced in
ice-cold isotonic homogenization buffer as previously
described [22,24,25].
Rat liver large granule and cytosolic fractions were iso-
lated by differential centrifugation as previously described
[43–46]. Plasma membrane was prepared according to the
method of Neville [47] as described by Authier et al.
[43,48,49]. The endosomal fraction was isolated by discon-
tinuous sucrose gradient centrifugation and collected at the
0.25–1.0 m sucrose interface [22,24,25]. Endosomal fractions
revealed no significant enrichment of lysosomal enzyme
markers (N-acetyl-b-d-glucosaminidase, relative specific
activity = 1.5; acid phosphatase, relative specific activ-
ity = 2.2), with the yield of enzymes accounting for
< 0.2% of that of the homogenate. The recovery of orga-
nelle enzyme markers in the nonsedimentable cytosolic frac-
tion was very low, and is in agreement with our previously

published biochemical characterizations [43,50].
Cell-free proteolysis and translocation of
endosome-associated ETA
Endosomal fractions isolated 30 min after the injection of
native ETA (15 lg per 100 g body weight) were suspended
at 1 mgÆmL
)1
in 0.15 m KCl, 5 mm MgCl
2
and 25 mm
Hepes (pH 7) or 25 mm citrate ⁄ phosphate buffer (pH 5–6)
in the presence or absence of 10 mm ATP and 0.01 lm
bafilomycin-A1. Samples were incubated at 37 °C for vari-
ous periods and subjected to reducing SDS ⁄ PAGE followed
by western blotting to determine the endosomal content
and integrity of ETA and ETA-A.
To specifically assess the membrane translocation of
intact and processed ETA through the endosomal mem-
brane, incubation mixtures were centrifuged for 60 min at
100 000 g. Pelleted endosomes and supernatants were then
subjected to reducing SDS ⁄ PAGE followed by western blot
analysis with antibody against ETA.
Cell-free translocation of mitochondria-associated
cytochrome c
A rat liver mitochondrial fraction (large-granule fraction)
was isolated by differential centrifugation as previously
described [43,46], and then resuspended at 7.5 mgÆmL
)1
in 0.15 m KCl, 5 mm MgCl
2

,1mm EDTA and 10 mm
Hepes (pH 7.5) in the presence or absence of native or
cathepsin-treated ETA, according to Uren et al. [51]. In
some experiments, KCl was omitted from the incubation
medium to disrupt mitochondria by hypotonic lysis. Sam-
ples were incubated at 37 °C for various periods (5 min to
2 h) and centrifuged for 15 min at 30 000 g. Supernatants
were subjected to reducing SDS ⁄ PAGE followed by western
blot analysis with antibody against cytochrome c.
Immunoblot analysis
Electrophoresed samples were transferred onto nitrocellu-
lose membranes for 60 min at 380 mA in transfer buffer
containing 25 mm Tris base and 192 mm glycine. The mem-
branes were blocked by a 3 h incubation with 5% skimmed
milk in 10 mm Tris ⁄ HCl (pH 7.5), 300 mm NaCl and
0.05% Tween-20. The membranes were then incubated with
primary antibody [mouse IgG against rat cytochrome c
(diluted 1 : 1000), mouse monoclonal antibody against rat
EEA1 (diluted 1 : 1000), rabbit polyclonal IgG against
either ETA (diluted 1 : 60 000), KX
5
KDEL (diluted
1 : 100) [22] or human EF-2 (diluted 1 : 500)] in the above
buffer for 16 h at 4 °C. The blots were then washed three
times with 0.5% skimmed milk in 10 mm Tris ⁄ HCl
(pH 7.5), 300 mm NaCl and 0.05% Tween-20 over a period
of 1 h at room temperature. The bound immunoglobulin
was detected with horseradish peroxidase-conjugated goat
anti-(rabbit IgG) or goat anti-(mouse IgG).
In vitro proteolysis of ETA by hepatic endosomes

and proteases
The endosomal fraction ( 1 lg) was incubated for varying
lengths of time at 37 °C with 10 lg of native ETA in 30 l L
of 25 mm citrate ⁄ phosphate buffer (pH 5) or 25 mm Hepes
buffer (pH 7) containing 6 mm CaCl
2
, in the presence or
absence of protease inhibitors. To determine the integrity
of ETA, the proteolytic reaction was stopped by the addi-
tion of reducing SDS ⁄ PAGE sample buffer, and this was
followed by SDS ⁄ PAGE and western blot analysis.
For some experiments, ETA (10 lg) was digested in vitro
for varying lengths of time with bovine cathepsin B or
cathepsin D (5 UÆmL
)1
Æmg
)1
)in50mm citrate ⁄ phosphate
buffer (pH 4–6), or human furin (10 UÆmL
)1
Æmg
)1
)in
50 mm Hepes buffer (pH 7) containing 10 mm CaCl
2
and
10 mm dithiothreitol. The proteolytic reaction was stopped
by the addition of reducing SDS ⁄ PAGE buffer, and this
was followed by SDS ⁄ PAGE and Coomassie Brilliant Blue
staining or western blot analysis.

ETA-catalyzed ADP-ribosylation of cytosolic EF-2
Native ETA (10 lg) was incubated with 5 UÆmL
)1
Æmg
)1
cathepsins B and D at 37 °C for 30 min in 25 mm
citrate ⁄ phosphate buffer (pH 4–6) and 0.1 m dithiothreitol
in the presence or absence of 5 lgÆmL
)1
PA and 1 lm E64.
Proteolysis of ETA in rat hepatic endosomes T. El Hage et al.
3746 FEBS Journal 277 (2010) 3735–3749 ª 2010 The Authors Journal compilation ª 2010 FEBS
The pretreated ETA was then neutralized with 0.5 m Hepes
(pH 7.4) and incubated with the cytosolic fraction (100 lg)
in 0.1 m Hepes (pH 7.4) containing 1 lm [
32
P]NAD for
15–60 min at 30 °C. The reaction was stopped by the
addition of Laemmli sample buffer, and this was followed
by SDS ⁄ PAGE and autoradiography.
Statistics
For some experiments, data were expressed as
mean ± standard deviation (SD). Statistical evaluations
were performed by Student’s t-test. Differences between
data groups were considered to be statistically significant if
the P-value was < 0.05.
Acknowledgements
We thank P. H. Cameron (McGill University, Mon-
treal, Canada) for reviewing the manuscript. We thank
N. Se

´
miramoth (Faculte
´
de Pharmacie, Chaˆ tenay-Mal-
abry, France) for assistance in the measurement of
caspase activity.
References
1 Koch C & Hoiby N (1993) Pathogenesis of cystic fibro-
sis. Lancet 341, 1065–1069.
2 Zdanovsky AG, Chiron M, Pastan I & FitzGerald DJ
(1993) Mechanism of action of Pseudomonas exotoxin.
Identification of a rate-limiting step. J Biol Chem 268,
21791–21799.
3 Jenkins CE, Swiatoniowski A, Issekutz AC & Lin T-J
(2004) Pseudomonas aeruginosa exotoxin A induces
human mast cell apoptosis by a caspase-8 and -3-depen-
dent mechanism. J Biol Chem 279, 37201–37207.
4 Kounnas MZ, Morris RE, Thompson MR, FitzGerald
DJ, Strickland DK & Saelinger CB (1992) The a 2 mac-
roglobulin receptor ⁄ low density lipoprotein receptor
binds and internalizes Pseudomonas exotoxin A. J Biol
Chem 267, 12420–12423.
5 Ogata M, Fryling CM, Pastan I & FitzGerald DJ
(1992) Cell-mediated cleavage of Pseudomonas exotoxin
between Arg279 and Gly280 generates the enzymatically
active fragment which translocates to the cytosol. J Biol
Chem 267, 25396–25401.
6 Theuer C, Kasturi S & Pastan I (1994) Domain II of
Pseudomonas exotoxin A arrests the transfer of translo-
cating nascent chains into mammalian microsomes.

Biochemistry 33, 5894–5900.
7 Smith DC, Spooner RA, Watson PD, Murray JL,
Hodge TW, Amessou M, Johannes L, Lord JM & Rob-
erts LM (2006) Internalized Pseudomonas exotoxin A
can exploit multiple pathways to reach the endoplasmic
reticulum. Traffic 7, 379–393.
8 Rasper DM & Merrill AR (1994) Evidence for the mod-
ulation of Pseudomonas aeruginosa exotoxin A-induced
pore formation by membrane surface charge density.
Biochemistry 33, 12981–12989.
9 Taupiac MP, Alami M & Beaumelle B (1996) Translo-
cation of full-length Pseudomonas exotoxin from endo-
somes is driven by ATP hydrolysis but requires prior
exposure to acidic pH. J Biol Chem 271, 26170–26173.
10 Alami M, Taupiac MP, Reggio H, Bienvenue A &
Beaumelle B (1998) Involvement of ATP-dependent
Pseudomonas exotoxin translocation from a late recy-
cling compartment in lymphocyte intoxication proce-
dure. Mol Biol Cell 9 , 387–402.
11 Jackson ME, Simpson JC, Girod A, Pepperkok R,
Roberts LM & Lord JM (1999) The KDEL retrieval
system is exploited by Pseudomonas exotoxin A, but
not by Shiga-like toxin-1, during retrograde transport
from the Golgi complex to the endoplasmic reticulum.
J Cell Sci 112, 467–475.
12 Seetharam S, Chaudhary VK, FitzGerald D & Pastan I
(1991) Increased cytotoxic activity of Pseudomonas exo-
toxin and two chimeric toxins ending in KDEL. J Biol
Chem 266, 17376–17381.
13 Inocencio NM, Moehring JM & Moehring TJ (1994)

Furin activates Pseudomonas exotoxin A by specific
cleavage in vivo and in vitro. J Biol Chem 269, 31831–
31835.
14 McKee ML & FitzGerald DJ (1999) Reduction of
furin-nicked Pseudomonas exotoxin A: an unfolding
story.
Biochemistry 38, 16507–16513.
15 Hessler JL & Kreitman R (1997) An early step in Pseu-
domonas exotoxin action is removal of the terminal
lysine residue, which allows binding to the KDEL
receptor. Biochemistry 36, 14577–14582.
16 Ogata M, Chaudhary VK, Pastan I & FitzGerald DJ
(1990) Processing of Pseudomonas exotoxin by a cellular
protease results in the generation of a 37,000-Da toxin
fragment that is translocated to the cytosol. J Biol
Chem 265, 20678–20685.
17 Moehring JM, Inocencio NM, Robertson BJ & Moeh-
ring TJ (1993) Expression of mouse furin in a Chinese
hamster cell resistant to Pseudomonas exotoxin A and
virus complements the genetic lesion. J Biol Chem 268,
2590–2594.
18 Pastan I, Chaudhary V & FitzGerald DJ (1992) Recom-
binant toxins as novel therapeutic agents. Annu Rev
Biochem 61, 331–354.
19 Morlon-Guyot J, Me
´
re
´
J, Bonhoure A & Beaumelle B
(2009) Processing of Pseudomonas aeruginosa exotoxin

is dispensable for cell intoxication. Infect Immun 77,
3090–3099.
20 Lund H, Takahashi K, Hamilton RL & Havel RJ
(1989) Lipoprotein binding and endosomal itinerary of
the low density lipoprotein receptor-related protein in
rat liver. Proc Natl Acad Sci USA 86, 9318–9322.
T. El Hage et al. Proteolysis of ETA in rat hepatic endosomes
FEBS Journal 277 (2010) 3735–3749 ª 2010 The Authors Journal compilation ª 2010 FEBS 3747
21 Herz J & Strickland DK (2001) LRP: a multifunctional
scavenger and signaling receptor. J Clin Invest 108,
779–784.
22 El Hage T, Merlen C, Fabrega S & Authier F (2007)
Role of receptor-mediated endocytosis, endosomal acid-
ification and cathepsin D in cholera toxin cytotoxicity.
FEBS J 274, 2614–2629.
23 Plotkowski MC, Po
´
voa HC, Zahm JM, Lizard G,
Pereira GM, Tournier JM & Puchelle E (2002) Early
mitochondrial dysfunction, superoxide anion produc-
tion, and DNA degradation are associated with non-
apoptotic death of human airway epithelial cells
induced by Pseudomonas aeruginosa exotoxin A. Am J
Respir Cell Mol Biol 26, 617–626.
24 El Hage T, Decottignies P & Authier F (2008) Endoso-
mal proteolysis of diphtheria toxin without toxin trans-
location into the cytosol of rat liver in vivo. FEBS J
275, 1708–1722.
25 Merlen C, Fayol-Messaoudi D, Fabrega S, El Hage T,
Servin A & Authier F (2005) Proteolytic activation of

internalized cholera toxin within hepatic endosomes by
cathepsin D. FEBS J 272, 4385–4397.
26 Fre
´
noy JP, Turpin E, Janicot M, Gehin-Fouque F &
Desbuquois B (1992) Uptake of injected 125I-ricin by
rat liver in vivo. Biochem J 284, 249–257.
27 Pavlovskis OR & Shackelford AH (1974) Pseudomonas
aeruginosa exotoxin in mice: localization and effect on
protein synthesis. Infect Immun 9 , 540–546.
28 Schafer W, Stroh A, Berghofer S, Seiler J, Vey M,
Kruse ML, Kern HF, Klenk HD & Garten W (1995)
Two independent targeting signals in the cytoplasmic
domain determine trans-Golgi network localization and
endosomal trafficking of the proprotein convertase
furin. EMBO J 14, 2424–2435.
29 Wouters S, Leruth M, Decroly E, Vandenbranden M,
Creemers JWM, van de Loo JHP, Ruysschaert JM &
Courtoy PJ (1998) Furin and proprotein convertase 7
(PC7) ⁄ lymphoma PC endogenously expressed in rat
liver can be resolved into distinct post-Golgi compart-
ments. Biochem J 336, 311–316.
30 Authier F, Mort JS, Bell AW, Posner BI & Bergeron
JJM (1995) Proteolysis of glucagon within hepatic endo-
somes by membrane-associated cathepsins B and D. J
Biol Chem 270, 15798–15807.
31 Legaard PK, LeGrand RD & Misfeldt ML (1992) Lym-
phoproliferative activity of Pseudomonas exotoxin A is
dependent on intracellular processing and is associated
with the carboxyl-terminal portion. Infect Immun 60,

1273–1278.
32 Pillay CS, Elliott E & Dennison C (2002) Endolysoso-
mal proteolysis and its regulation. Biochem J 363, 417–
429.
33 Corboy MJ & Draper RK (1997) Elevation of vacuolar
pH inhibits the cytotoxic activity of furin-cleaved exo-
toxin A. Infect Immun 65, 2240–2242.
34 Chiron MF, Fryling CM & FitzGerald DJ (1994)
Cleavage of pseudomonas exotoxin and diphtheria toxin
by a furin-like enzyme prepared from beef liver. J Biol
Chem 269, 18167–18176.
35 Molloy SS, Bresnahan PA, Leppla SH, Klimpel KR &
Thomas G (1992) Human furin is a calcium-dependent
serine endoprotease that recognizes the sequence
Arg-X-X-Arg and efficiently cleaves anthrax toxin
protective antigen. J Biol Chem 267, 16396–16402.
36 Hotchkiss R, Dunne WM, Swanson PE, Davis CG,
Tinsley KW, Chang KC, Buchman TG & Karl IE
(2001) Role of apoptosis in Pseudomonas aeruginosa
pneumonia. Science 294, 1783–1791.
37 Rajan S, Cacalano G, Bryan R, Ratner AJ, Sontich
CH, van Heerckeren A, Davis P & Prince A (2000)
Peudomonas aeruginosa induction of apoptosis in respi-
ratory epithelial cells: analysis of the effects of cystic
fibrosis transmembrane conductance regulator dysfunc-
tion and bacterial virulence factor. Am J Respir Cell
Mol Biol 23, 304–312.
38 Jendrossek V, Grassme
´
H, Mueller I, Lang F &

Gulbins E (2001) Pseudomonas aeruginosa-induced
apoptosis involves mitochondria and stress-activated
protein kinases. Infect Immun 69, 2675–2683.
39 Hauser AR & Engel JN (1999) Pseudomonas aeruginosa
induces type-III-secretion-mediated apoptosis of
macrophages and epithelial cells. Infect Immun 67,
5530–5537.
40 Lowry OH, Rosebrough NJ, Farr AL & Randall RJ
(1951) Protein measurement with the Folin phenol
reagent. J Biol Chem 193, 265–275.
41 Touster O, Aronson NN, Dulaney JT & Hendrickson
H (1970) Isolation of rat liver plasma membranes. Use
of nucleotide pyrophosphate and phosphodiesterase I as
marker enzymes. J Cell Biol 47, 604–618.
42 Trouet A (1974) Isolation of modified liver lysosomes.
Methods Enzymol 31, 323–329.
43 Authier F, Janicot M, Lederer F & Desbuquois B
(1990) Fate of injected glucagon taken up by rat liver
in vivo. Degradation of internalized ligand in the
endosomal compartment. Biochem J 272, 703–712.
44 Authier F & Desbuquois B (1991) Degradation of glu-
cagon in isolated liver endosomes. Biochem J 280, 211–
218.
45 Authier F, Rachubinski RA, Posner BI & Bergeron
JJM (1994) Endosomal proteolysis of insulin by an
acidic thiol metalloprotease unrelated to insulin-degrad-
ing enzyme. J Biol Chem 269, 3010–3016.
46 Authier F, Cameron PH, Merlen C, Kouach M & Bri-
and G (2003) Endosomal proteolysis of glucagon at
neutral pH generates the bioactive degradation product

miniglucagon-(19–29). Endocrinology 144, 5353–5364.
47 Neville DM (1968) Isolation of an organ specific protein
antigen from cell surface membrane of rat liver. Biochim
Biophys Acta 154, 540–552.
Proteolysis of ETA in rat hepatic endosomes T. El Hage et al.
3748 FEBS Journal 277 (2010) 3735–3749 ª 2010 The Authors Journal compilation ª 2010 FEBS
48 Authier F, Desbuquois B & De Galle B (1992) Ligand-
mediated internalization of glucagon receptors in intact
rat liver. Endocrinology 131, 447–457.
49 Authier F, Di Guglielmo GM, Danielsen GM & Bergeron
JJM (1998) Uptake and metabolic fate of [His
A8
,
His
B4
,Glu
B10
,His
B27
]insulin in rat liver in vivo. Biochem J
332, 421–430.
50 Authier F, Me
´
tioui M, Fabrega S, Kouach M & Briand
G (2002) Endosomal proteolysis of internalized insulin
at the C-terminal region of the B chain by cathepsin D.
J Biol Chem 277, 9437–9446.
51 Uren RT, Dewson G, Bonzon C, Lithgow T,
Newmeyer DD & Kluck RM (2005) Mitochondrial
release of pro-apoptotic proteins. Electrostatic

interactions can hold cytochrome-c but not
Smac ⁄ Diablo to mitochondrial membranes. J Biol
Chem 280, 2266–2274.
T. El Hage et al. Proteolysis of ETA in rat hepatic endosomes
FEBS Journal 277 (2010) 3735–3749 ª 2010 The Authors Journal compilation ª 2010 FEBS 3749