Báo cáo khoa học: Puroindoline-a and a1-purothionin form ion channels in giant liposomes but exert different toxic actions on murine cells pptx

Puroindoline-a and a1-purothionin form ion channels
in giant liposomes but exert different toxic actions
on murine cells
Paola Llanos
1
, Mauricio Henriquez
1
, Jasmina Minic
2
, Khalil Elmorjani
3
, Didier Marion
3
,
Gloria Riquelme
1
, Jordi Molgo
´
2
and Evelyne Benoit
2
1 Instituto de Ciencias Biome
´
dicas, Facultad de Medicina, Universidad de Chile, Santiago, Chile
2 Laboratoire de Neurobiologie Cellulaire et Mole
´
culaire, UPR 9040, Centre National de la Recherche Scientifique, Gif-sur-Yvette cedex,
France
3 Biopolyme
`
res Interactions Assemblages, Institut National de la Recherche Agronomique, Nantes, France

Various cationic lipid-binding proteins, the folding of
which is stabilized by four or five disulfide bonds, have
been isolated from wheat endosperm. They include
lipid-transfer proteins (LTPs), puroindolines (PINs)
and a⁄ b-purothionins (PTHs) [1]. PINs are restricted
to Triticae and Avenae species [2,3], whereas LTPs and
PTHs are ubiquitous plant proteins found in most
plant organs [1,4,5]. PIN-a and PIN-b, two isoforms
sharing  60% sequence homology, have been purified
from wheat seeds. PIN-a displays a unique trypto-
phan-rich domain (WRWKWWK), which is slightly
truncated in PIN-b (WPTKWWK). The 3D structures
of LTPs and PINs are closely related and rich in
a-helices, as suggested by their cysteine pairing and
secondary-structure characterization [1,6], and they dif-
fer from the structure of the PTHs [5]. Furthermore, in
contrast with LTPs, PINs and PTHs can be isolated
by Triton X-114 phase partitioning [7], an observation
that is in agreement with differences in their lipid-
binding properties. Indeed, whereas LTPs bind lipid
monomers in a hydrophobic cavity, PINs and PTHs
interact with lipid aggregates, e.g. micelles and lipid
bilayers [1].
Because of their toxic activity against fungi, yeast
and bacteria, PTHs have been suggested to play a role
in plant defence against microbial pathogens [4,5].
PINs are also thought to have a role in plant defence
because of their antifungal properties in vitro, and
especially because they enhance the antimicrobial
effects of PTHs [8]. In addition, leaf extracts of trans-

genic rice plants expressing genes encoding PINs (pinA
and ⁄ or pinB) reduce in vitro the growth of rice fungal
Keywords
a1-purothionin; giant liposomes; ion
channels; neuromuscular transmission;
puroindoline-a
Correspondence
E. Benoit, Laboratoire de Neurobiologie
Cellulaire et Mole
´
culaire, UPR 9040, Centre
National de la Recherche Scientifique, ba
ˆ
t.
32–33, 91198 Gif-sur-Yvette cedex, France
Fax: +33 169 82 41 41
Tel: +33 169 82 36 52
E-mail:
(Received 20 July 2005, revised 13 February
2006, accepted 17 February 2006)
doi:10.1111/j.1742-4658.2006.05185.x
Puroindoline-a (PIN-a) and a1-purothionin (a1-PTH), isolated from wheat
endosperm of Triticum aestivum sp., have been suggested to play a role in
plant defence mechanisms against phytopathogenic organisms. We investi-
gated their ability to form pores when incorporated into giant liposomes
using the patch-clamp technique. PIN-a formed cationic channels
( 15 pS) with the following selectivity K
+
>Na
+

 Cl
–
. Also, a1-PTH
formed channels of  46 pS and 125 pS at +100 mV, the selectivity of
which was Ca
2+
>Na
+
 K
+
 Cl
–
and Cl
–
 Na
+
, respectively. In
isolated mouse neuromuscular preparations, a1-PTH induced muscle mem-
brane depolarization, leading to blockade of synaptic transmission and
directly elicited muscle twitches. Also, a1-PTH caused swelling of differen-
tiated neuroblastoma NG108-15 cells, membrane bleb formation, and dis-
organization of F-actin. In contrast, similar concentrations of PIN-a had
no detectable effects. The cytotoxic actions of a1-PTH on mammalian cells
may be explained by its ability to induce cationic-selective channels.
Abbreviations
EPP, endplate potential; LTP, lipid-transfer protein; MEPP, miniature endplate potential; PIN, puroindoline; PTH, purothionin.
1710 FEBS Journal 273 (2006) 1710–1722 ª 2006 The Authors Journal compilation ª 2006 FEBS
pathogens [9]. The generalized toxicity of PTHs may
be due to their ability to form ion channels in the
membrane of target cells, resulting in dissipation of

ion concentration gradients essential for the mainten-
ance of cellular homoeostasis [10–13]. Also, b-PTH
extracted from wheat flour has been shown to form
cation-selective ion channels in artificial lipid bilayer
membranes and in the plasmalemma of rat hippocam-
pal neurons [14]. PIN-a and a1-PTH have also been
reported to swell the nodes of Ranvier of frog myeli-
nated axons, and pore formation in the nodal mem-
brane has been suggested to be responsible for these
effects [15]. In addition, PINs have also been shown to
be cytotoxic to Xenopus oocytes [16]. However, the
mechanisms involved in the toxicity of PTHs and PINs
to mammalian cells remain poorly understood. There-
fore, as a first step toward understanding these mecha-
nisms, we characterized (a) the pore-forming activity
of PIN-a and a1-PTH in giant liposomes and (b) their
toxicity to mammalian phrenic nerve ⁄ hemidiaphragm
muscle preparations and cultured neuroblastoma
(NG108-15) cells. A preliminary account of part of this
work has been published in abstract form [17].
Results
Molecular masses of purified PIN-a and a1-PTH
A typical electrospray mass spectrum of purified wheat
PIN-a (Fig. 1A) reveals that its apparent heterogeneity
is related to complex post-translational proteolytic
maturation which leads to two major forms (M
r
12 750 and 12 919) and three minor ones (M
r
12 634.7, 12 803.5 and 13 083.6). However, as reported

by Blochet et al. [18], all these polypeptides originate
from a unique polypeptide template with different
extensions at both the N-terminus and C-terminus
(Fig. 1A). The mass spectrum of a1-PTH is depicted in
Fig. 1B. All the masses reported here fit very well with
the expected calculated molecular masses for native
PIN-a and a1-PTH.
PIN-a forms ionic channels in giant liposomes
Seals of high-resistance and excised patches in an
‘inside out’ configuration were obtained from 19 prep-
arations of giant liposomes containing PIN-a. Forty-
five of 72 patches studied ( 60%) exhibited channel
activity. Usually, multiple current levels corresponding
to a similar conductance were observed, and at least
two distinct levels were detected in 35 of the 45 pat-
ches ( 78%). The unitary level was difficult to
observe in isolation, which suggested the presence of
substrates or clustering of channels in the patch. How-
ever, because it was not possible to distinguish between
these two possibilities, we assume that each level above
the baseline corresponds to a single channel, which
opens and closes independently. Unitary currents
recorded at a holding potential of +100 mV are
shown in Fig. 2A, and the corresponding current
amplitude distribution is depicted in Fig. 2B. Six chan-
nels were present in the patch, as judged by the num-
ber of simultaneous unitary current steps and
histogram peaks. A unitary conductance of 14.8 ±
0.6 pS (n ¼ 17) was determined, between )80 and
+80 mV, from the slope of the current–voltage rela-

tionship (Fig. 2C).
4,700
4,800
4,900 5,000 5,100 5,200
α1-PTH
4,921
B
12,60012,200 13,000 13,400 13,800
DVA-(PIN)
-GTIG
12,919
DVA-(PIN)
-GTIGY
13,083.6
VA-(PIN)
-GT
12,634.7
VA-(PIN)
-GTIG
12,803.05
DVA-(PIN)
-GT
12,750
A
Fig. 1. MALDI-TOF mass spectrum of PIN-a and a1-PTH. Deconvo-
luted and reconstructed electrospray mass spectra from multi-
charged ion spectra of the purified PIN-a (A) and a1-PTH (B). Note
the homogeneity of the protein preparations.
P. Llanos et al. Toxic actions of puroindoline-a and a1-purothionin
FEBS Journal 273 (2006) 1710–1722 ª 2006 The Authors Journal compilation ª 2006 FEBS 1711

The selectivity of the channels for Na
+
vs. Cl
–
was
determined by increasing or decreasing the NaCl con-
centration in the bath solution. In the presence of a
high NaCl concentration (440 mm instead of 140 mm)
in the bath, the reversal potential of the current recor-
ded in response to potential ramps shifted from 0 to
)25 ± 1 mV (n ¼ 3, Fig. 2D), which is close to the
)28 mV theoretical equilibrium potential for Na
+
(the
equilibrium potential for Cl
–
was +29.6 mV under this
condition). Similarly, in the presence of a low NaCl
concentration (40 mm instead of 140 mm) in the bath,
the reversal potential of the current shifted from 0 to
+24 mV (data not shown), which is close to the
+30 mV equilibrium potential for Na
+
. The Na
+
to
Cl
–
permeability ratio (P
Na

⁄ P
Cl
) was  13 (n ¼ 5). To
determine the K
+
to Na
+
permeability ratio
(P
K
⁄ P
Na
), we replaced the bath NaCl with KCl. When
140 mm KCl was perfused in the bath solution, the
current recorded in response to a potential ramp
showed an almost linear current–voltage relation-
ship, and it had a reversal potential of )9.2 ± 0.8 mV
(n ¼ 8). Under these conditions, the permeability ratio
was 1.43 ± 0.04 (n ¼ 8). These results indicate that
PIN-a forms a cationic channel, the permeability
sequence of which is K
+
>Na
+
 Cl
–
.
a1-PTH forms ionic channels in giant liposomes
Giant liposomes containing a1-PTH also produced
excised patches with seals of high resistance. Channel

activity was found in 31 ( 41%) of 75 recorded pat-
ches. Single channels with a high unitary conductance
and single channels with a low unitary conductance
were detected in 32% (n ¼ 10) and 68% (n ¼ 21) of
the recordings, respectively. In two independent experi-
ments, low-conductance and high-conductance open-
ings were detected simultaneously in the same patch,
but these data have not been included in this study
because of difficulties with their analysis.
Typical current recordings through high-conduct-
ance channels formed by a1-PTH are shown in
Fig. 3A. The unitary conductance was 125 and 100 pS
at holding potentials of +100 and )100 mV, respect-
ively. Figure 3B depicts the current vs. voltage plot at
different holding potentials in the presence of 140 mm
and 40 mm NaCl in the bath solution. When the bath
concentration of NaCl was decreased, the reversal
potential of the current shifted from 0 to )20 mV,
which is close to the )28 mV equilibrium potential
for Cl
–
. The calculated Cl
–
to Na
+
permeability ratio
(P
Cl
⁄ P
Na

) was  7, which indicates that the high-
500 ms
1 pA
500 ms
1 pA
5 s
1 pA
AB
V (mV)
-100 -50 50 100
-1.5
-1.0
-0.5
0.5
1.0
1.5
C
I (pA)
100-100
-50
50
-100
-50
50
100
V (mV)
D
I (pA)
I (pA)
%emiT

0246
1.5
Fig. 2. Ion-channel activities exhibited by
giant liposomes containing PIN-a. (A) Unitary
current traces recorded at a holding potent-
ial of +100 mV. The zero current level is
indicated by the dotted line, and channel
openings are indicated by upward deflect-
ions. The arrows show in an expanded time
base the corresponding unitary currents.
(B) Time distribution of current amplitude
corresponding to the recordings shown in
(A). The time was expressed as a percent-
age of the total recording time. (C) Current–
voltage relationship obtained from current
amplitude distributions at various holding
potentials. A unitary conductance of
14.8 ± 0.6 pS (n ¼ 17) was determined
from the slope of the relationship by linear
regression between )80 and +80 mV. (D)
Representative currents recorded during
potential ramps from )100 to +100 mV in
the presence of either 140 or 440 m
M NaCl
in the bathing solution. Under these condit-
ions, the voltages corresponding to zero
current were 0 and )24.3 mV (arrow),
respectively.
Toxic actions of puroindoline-a and a1-purothionin P. Llanos et al.
1712 FEBS Journal 273 (2006) 1710–1722 ª 2006 The Authors Journal compilation ª 2006 FEBS

conductance channel formed by a1-PTH is an anionic
channel. The Ca
2+
selectivity of the channels was nil.
Indeed, when the concentration of CaCl
2
was increased
from 2.6 mm to 10 or 20 mm in the bath solution, no
significant effect on current amplitude or reversal
potential values was detected in response to potential
ramps (data not shown).
Currents through low-conductance channels formed
by a1-PTH were recorded at holding potentials varying
from 0 to ± 80 mV in steps of 40 mV (Fig. 4A). Unit-
ary conductances of 46 ± 5 and 34 ± 2 pS were cal-
culated from current potential relationships at holding
potentials of +100 and )100 mV, respectively
(Fig. 4B), observed during 21 experiments using sym-
metrical NaCl (140 mm, n ¼ 18) or sodium gluconate
(140 mm, n ¼ 3) concentrations. When the NaCl
concentration in the bath solution was decreased from
140 to 40 mm, the reversal potential of the current
recorded in response to potential ramps shifted from 0
to +23 mV (Fig. 4C), which is close to the Na
+
equi-
librium potential (+ 27.8 mV). A P
Na
⁄ P
Cl

of 11 was
calculated. The low-conductance channels were almost
equally selective to Na
+
and K
+
. The P
K
⁄ P
Na
was
A
-60
-30
0
30
60
1 s
150 mV
50 mV
0 mV
-150 mV
-50 mV
B
I ( pA)
V (mV)
-150 -100
50 100 150
-20
-15

-10
-5
5
10
15
20
25
-20 mV
I (pA)
Fig. 3. High-conductance channel activities exhibited by giant lipo-
somes containing a1-PTH. (A) Unitary current traces recorded at
the indicated holding potentials. (B) Current–voltage relationships in
the presence of either 140 m
M NaCl (s)or40mM NaCl (d) in the
bathing solution. Under these conditions, the voltages correspond-
ing to zero current were 0 and )20 mV (arrow), respectively.
A
-20
0
20
1 s
80 mV
40 mV
0 mV
-80 mV
-40 mV
B
V (mV)
I (pA)
-150 -100 -50 50 100 150

-4
-2
2
4
6
8
23 mV
V (mV)
I (pA)
-100 -50 50 100
-100
-50
50
100
C
Fig. 4. Low-conductance channel activities exhibited by giant lipo-
somes containing a1-PTH. (A) Unitary current traces recorded in
response to holding potentials varying from 0 to ± 80 in steps of
40 mV. (B) Representative current–voltage relationship. The calcula-
ted unitary conductance was 51 and 35 pS at holding potentials of
+100 mV and )100 mV, respectively. (C) Representative currents
recorded during potential ramps from )100 mV to +100 mV in the
presence of either 140 or 40 m
M NaCl in the bathing solution.
Under these conditions, the voltages corresponding to zero current
were 0 and +23 mV (arrow), respectively.
P. Llanos et al. Toxic actions of puroindoline-a and a1-purothionin
FEBS Journal 273 (2006) 1710–1722 ª 2006 The Authors Journal compilation ª 2006 FEBS 1713
1.10 ± 0.04 (n ¼ 4), as calculated from changes in
the current reversal potential, e.g. from 0 to

)2.5 ± 1.1 mV (n ¼ 4), brought about by replacing
NaCl (140 mm) with KCl (140 mm) in the bath solu-
tion. These results indicate that the low-conductance
channel formed by a1-PTH is a cationic channel.
The selectivity of the low-conductance channel to
bivalent cations was studied by changing the CaCl
2
concentration in the bath solution. Figure 5A shows
unitary currents, recorded at a holding potential of
0 mV, in the presence of 2.6 mm (control conditions)
and 20 m m CaCl
2
. In response to potential ramps,
the reversal potential shifted from 0 (Fig. 5B) to
)8.0 ± 0.8 mV (n ¼ 4, Fig. 5C) when the CaCl
2
con-
centration was increased from 2.6 to 10 mm, and it
was )13.3 ± 0.4 mV (n ¼ 5) when the CaCl
2
concen-
tration was 20 mm. Under these conditions, the expec-
ted equilibrium potential calculated for Ca
2+
was
)17.3 mV and )26.2 mV for 10 and 20 mm CaCl
2
,
respectively. A Ca
2+

to Na
+
permeability ratio
(P
Ca
⁄ P
Na
) of 5 was calculated from the changes in the
measured reversal potential. Thus, the relative ionic
permeability sequence for the cationic channel formed
by a1-PTH is Ca
2+
>Na
+
 K
+
 Cl
–
.
Effects of a1-PTH and PIN-a on isolated mouse
neuromuscular preparations
The addition of a1-PTH (0.01–1 lm) to the physiologi-
cal medium bathing isolated preparations produced a
concentration-dependent decrease in muscle twitches
and tetanic responses evoked by nerve stimulation at
0.2 and 40 Hz, respectively (Fig. 6A). The concentra-
tion of a1-PTH that reduced the contraction amplitude
by 50% was 0.16 lm (Fig. 6B). Complete blockade of
nerve-evoked muscle twitches and tetanic responses
occurred with 1 lm a1-PTH (n ¼ 10), and the block-

ade was not reversed after extensive washing with the
standard physiological solution. Similar concentrations
of a1-PTH also blocked twitches evoked by direct
electric stimulation of the muscle (Fig. 6A,B). Thus,
a1-PTH is toxic to isolated mouse phrenic nerve ⁄ hemi-
diaphragm muscle preparations. In contrast, when we
examined the ability of PIN-a (0.01–1 lm) to alter
muscle twitches and tetanic responses evoked by nerve
stimulation at 0.2 and 40 Hz, respectively, no signifi-
cant changes were detected in the contraction ampli-
tude (Fig. 6B).
Membrane permeability changes caused by the pore-
forming ability of a1-PTH may explain the above
effects. Therefore, we performed intracellular record-
ings to measure the effect of a1-PTH and PIN-a on
the resting membrane potential of mouse hemidia-
phragm muscle fibres. When added to the standard
medium, a1-PTH (0.05–1 lm) caused dose-dependent
membrane depolarization (Fig. 7A). A representative
recording of the time course of 1 lm a1-PTH-induced
depolarization of skeletal muscle fibres is shown in
Fig. 7B. The time required by a1-PTH to exert half-
010 30 40 50 60
4 pA
0 pA
5 s
4 pA
0 pA
Control
CaCl

2
(20mM in bath)
0 mV
min
A
V (mV)
I (pA)
-100 -50 50 100
-100
-50
50
100
-100 50 100
-100
-50
50
100
I (pA)
V (mV)
CaCl
2
(10mM in bath)
B
C
Fig. 5. Low-conductance channel selectivity for Ca
2+
. (A) Current
traces recorded at a holding potential of 0 mV when the bath con-
centration of CaCl
2

was increased from 2.6 mM (control) to 20 mM
(arrow). The dotted lines indicate the zero current level. The arrows
show in an expanded time basis unitary currents. (B,C) Representa-
tive currents recorded during the same experiment in response to
potential ramps from )100 mV to +100 mV in the presence of
either 2.6 m
M (B) or 10 mM (C) CaCl
2
in the bathing solution. Under
these conditions, the voltages corresponding to zero current were
0 (B) and )9 mV (C, arrow).
Toxic actions of puroindoline-a and a1-purothionin P. Llanos et al.
1714 FEBS Journal 273 (2006) 1710–1722 ª 2006 The Authors Journal compilation ª 2006 FEBS
maximal depolarization was 3.5 ± 0.8 min (n ¼ 4).
The magnitude of the depolarization was independent
of the external CaCl
2
concentration between 0 and
2mm (Fig. 7C). However, when a1-PTH was added to
the standard medium in which the CaCl
2
concentration
was raised from 2 mm to 5 or 10 mm, no significant
change was detected in the resting membrane potential
of the muscle fibres (Fig. 7B,C). In contrast with the
marked effect of a1-PTH, PIN-a (0.05–1 lm) did not
significantly alter the resting membrane potential of
the muscle fibres (Fig. 7A). However, at a higher con-
centration (10 lm), the protein hyperpolarized the
muscle membrane by 21 ± 2.5 mV within about 5 min

(n ¼ 4).
Analysis of synaptic transmission at single neuro-
muscular junctions revealed that a1-PTH (0.05–1 lm)
produced a dose-dependent decrease in endplate poten-
tial (EPP) amplitude. Thus, in the presence of 0.25 lm
a1-PTH, EPPs had a subthreshold amplitude, being
unable to reach the threshold for action potential gen-
eration in muscle fibres, and were almost completely
blocked by 1 lm a1-PTH (Fig. 8A). Parallel recordings
of spontaneous miniature endplate potentials (MEPPs)
showed that 0.5 lm a1-PTH increased the frequency of
MEPPs. Thus, MEPP frequency was 0.7 ± 0.2 s
-1
(n ¼ 12) in control conditions and 6.7 ± 0.4 s
1
(n ¼ 6)
after the addition of 0.5 lm a1-PTH. In addition,
a1-PTH caused a marked decrease in MEPP ampli-
tude, which attained the basal noise level with 1 lm
a1-PTH (Fig. 8B). This precluded the recording of
A
B
Fig. 6. Effects of a1-PTH and PIN-a on nerve-evoked and directly
elicited muscle twitches. (A) Superimposed tracings of muscle twit-
ches evoked by nerve stimulation (left panel, 0.2 Hz), tetanic nerve
stimulation (middle panel, 40 Hz), and direct muscle stimulation
(right panel). Tracings were recorded before and after 20 min expo-
sure to a1-PTH (0.05–1 l
M). (B) Dose–response curves of the
effects of a1-PTH (circles) and PIN-a (squares) on nerve-evoked

(filled symbols), and directly elicited muscle twitch (s). The twitch
tension is expressed with respect to controls and as means ± SEM
for n experiments (numbers beside data points). Note the complete
blockade of the twitch response in the presence of 1 l
M a1-PTH,
and the quasi-absence of effect of similar concentrations of PIN-a.
Protein concentration
α1-PTH (2 mM CaCl
2
)
PIN-a (2 m
M CaCl
2
)
)Vm(laitnetopenarbmemgnitseR
-90
-80
-70
-60
-50
-40
-30
-20
-10
0
A
)Vm(laitnetop
enarbmemgn
itseR
Control

α1- PTH (1 µ
M)
-90
-80
-70
-60
-50
-40
-30
-20
-10
0
C
CaCl
2
concentration
0
-10
-20
-30
-40
-50
-60
-70
-80
0 1 2 3 4 5 6 7 29 30 31 32
Time (min)
)Vm(laitnetopenarbmemgnitseR
1 µM α1-PTH (2 mM CaCl
2

)
1 µ
M α1-PTH (10 mM CaCl
2
)
B
//
//
//
Fig. 7. Effects of a1-PTH and PIN-a on the resting membrane
potential of skeletal muscle fibres (A), and the influence of extracell-
ular Ca
2+
concentration on the effect of a1-PTH (B and C). Note
that in (A) only a1-PTH produces membrane depolarization, and in
(B) and (C) increasing extracellular Ca
2+
concentration (from 2 to
10 m
M) markedly reduces a1-PTH-induced muscle depolarization,
whereas decreasing extracellular Ca
2+
concentration (from 2 to
0m
M) has no significant effect. In (A) and (C), each column repre-
sents the mean ± SEM obtained from 3 to 31 fibres. In (B), the
points represent the membrane potential of single muscle fibres as
a function of time after addition of a1-PTH to the medium.
P. Llanos et al. Toxic actions of puroindoline-a and a1-purothionin
FEBS Journal 273 (2006) 1710–1722 ª 2006 The Authors Journal compilation ª 2006 FEBS 1715

MEPPs with concentrations higher than 0.5 lm
a1-PTH. The above results indicate that a1-PTH,
within the range of concentrations studied, causes per-
meability changes in the presynaptic and postsynaptic
membranes of the neuromuscular junction. In contrast,
MEPPs (Fig. 8C) and EPPs were not significantly
affected by 1 lm PIN-a.
Cytotoxic effects of a1-PTH and PIN-a on
neuroblastoma (NG108-15) cells
The effects of a1-PTH and PIN-a were studied on the
morphology of NG108-15 cells stained with the fluor-
escent dye FM1-43 and imaged with a confocal laser
scanning microscope (see Experimental procedures).
Images from each experiment were processed identic-
ally, and the effects were quantified using the same
cells examined before and during the action of the
proteins. The 3D projected area of cells was measured
as an index of cell volume. The addition of 10 lm
a1-PTH to the standard mammalian physiological
solution (e.g. containing 2 mm CaCl
2
) produced, after
a latent period of about 15 min, a marked increase in
the fluorescence intensity and a slight but significant
(P ¼ 0.0001) swelling of the cells (Fig. 9Ab). Within
30–45 min, the 3D projected area of the cells reached a
maximum increase of 128 ± 17% (n ¼ 14), when com-
pared with control values. In addition, large membrane
bleb formation followed by bleb dilation was consis-
tently observed, and, in most cases, the blebs attained

the size of the cells. A similar increase in both fluores-
cence intensity and 3D projected area occurred when
the cells were exposed to 10 lm a1-PTH in a CaCl
2
-
free medium. However, under these conditions, bleb
formation was not observed (Fig. 9Ad). In contrast,
exposure to 10 lm PIN-a had no detectable effect on
the morphology of NG108-15 cells (Fig. 9Bc),
although higher concentrations (50 and 100 lm) pro-
duced a 10–15% decrease in the cells’ 3D projected
area (Fig. 9Bb,c).
Effects of a1-PTH and PIN-a on the actin network
of NG108-15 cells
The significant membrane blebbing observed in the
presence of a1-PTH, but not in the presence of PIN-
a, prompted us to determine whether the two pro-
teins affect the cytoskeleton of NG108-15 cells. Thus,
immunofluorescence studies were performed to detect
eventual changes in the actin network organization
after exposure to a1-PTH (10 lm) or PIN-a (10 lm)
for 2–4 h. In comparison with control cells, and with
cells treated with PIN-a, significant changes in fila-
mentous actin immunolabelling distribution were
observed in cells treated with a1-PTH (Fig. 9C). Dis-
organization and disarray of actin were reflected by
disruption and clumping of actin filaments, which
was depicted as a punctuate pattern throughout the
cytoplasm (Fig. 9Cb). Similar results were observed
in NG108-15 cells after AlexaFluor-594-conjugated

phalloidin staining to visualize F-actin (data not
shown).
Discussion
In this work, the ion channels formed by PIN-a and
a1-PTH reconstituted into giant liposomes were
characterized, and their toxicity examined in murine
isolated neuromuscular preparations and cultured
NG108-15 cells. To our knowledge, this is the first
study to demonstrate that highly purified a1-PTH
0.5 mV
1.00 µM0.00 µM PIN-a
5 ms
C
0.25 µM0.00 µM
20 mV
1.00 µM
α1-PTH
5 ms
A
B
0.00 µM
0.5 mV
0.50 µM 1.00 µM
α1-PTH
5 ms
Fig. 8. Effects of a1-PTH and PIN-a on nerve-evoked action potent-
ial, EPPs and MEPPs recorded from isolated hemidiaphragm mus-
cles. (A) Single action potential recorded before (left trace), and
EPPs recorded after, 20 min exposure to 0.25 l
M (middle trace)

and 1 l
M (right trace) a1-PTH. The arrow indicates the stimulation
artefact of the phrenic nerve. (B) Average of 30 sequential MEPPs
recorded before (left trace) and after 20 min exposure to 0.5 l
M
(middle trace) and 1 lM a1-PTH (right trace). (C) Average of 30
sequential MEPPs recorded before (left trace) and after 20 min
exposure to 1 l
M PIN-a (right trace). Note the subthreshold EPP (A,
middle trace), the reduction and complete block of averaged
MEPPs induced by a1-PTH (B, right trace), and the absence of
effect of PIN-a on the amplitude of averaged MEPPs (C, right
trace). Note the different scales in A, B and C.
Toxic actions of puroindoline-a and a1-purothionin P. Llanos et al.
1716 FEBS Journal 273 (2006) 1710–1722 ª 2006 The Authors Journal compilation ª 2006 FEBS
forms ion channels in biological and artificial mem-
branes. In addition, we found that PIN-a forms a cat-
ion-selective channel with a 15 pS conductance. This
channel is 13 times more permeable to a univalent cat-
ion (Na
+
) than to Cl
–
and 1.4-fold more permeable to
K
+
than to Na
+
. These results considerably extend
those previously obtained on voltage-clamped Xenopus

oocytes [16].
In the case of a1-PTH, two types of channels with
different conductance and selectivity were detected.
One of them is an anion channel with a conductance
larger than 100 pS (high-conductance channel), which
is 7 times more permeable to Cl
–
than to Na
+
. Our
results constitute the first description of single-channel
anion currents induced by a1-PTH. The low-conduct-
ance channel formed by a1-PTH is a cationic channel
with single unitary conductance of 30–45 pS, which is
11 times more permeable to Na
+
than to Cl
–
, and has
a permeability ratio between Na
+
and K
+
close to 1.
However, this channel is 5 times more permeable to
Ca
2+
than to Na
+
. Thus, the cation selectivity order

for this channel is the following: Ca
2+
>Na
+
 K
+
.
These results are consistent with and extend previ-
ous observations suggesting that PTHs form cation-
selective ion channels, and in particular with data
obtained with b-PTH showing the formation of cation-
selective ion channels in artificial lipid bilayer mem-
branes and in the plasmalemma of rat hippocampal
neurons [14].
Although a concentration-dependent study of a1-
PTH has not been performed, it is possible that the
two different channel behaviours detected are the con-
sequence of protein–protein interactions in the recon-
stituted system, leading to channel clustering or similar
processes. Thus, it is likely that progressive recruitment
of additional monomers will contribute to increase the
pore size. The formation of transmembrane chan-
nels ⁄ pores by bundles of amphipathic a-helices of
a1-PTH and PIN-a polypeptides may occur via a ‘bar-
rel-stave’ mechanism [19], in such a manner that their
hydrophobic surfaces interact with the lipid core of the
membrane and their hydrophilic surfaces point inward,
producing an aqueous pore. According to this model,
10 µm
a

b
Control 50 µ
M
PIN-a
a
b
10 µm
10 µ
M
α1-PTH (0 m
M
CaCl
2
)
c
d
Control (0 m
M
CaCl
2
)
Control 10 µ
M
PIN-a10 µ
M
α1-PTH
* P > 0.02
* * P< 0.002
slle
c 51-80

1GN fo ae
ra evitaleR
12
*
*
20
* *
16
* *
16
* *
34
* *
* *
21
20
Concentration (time of application)
10 µ
M
(90 min)
50 µ
M
(30min)
50 µ
M
(60 min)
50 µ
M
(120 min)
100 µ

M
(15 min)
100 µ
M
(30 min)
100 µ
M
(90 min)
0.0
0.2
0.4
0.6
0.8
1.0
1.2
c
B
A
C
a
b
c
5 µm
Fig. 9. Effects of a1-PTH and PIN-a on NG108-15 cells. In (A) and
(B), the cells were stained with 2 l
M FM1-43 dye for 30 min, and
thereafter abundantly washed with dye-free solution before imaging
and the addition of a1-PTH or PIN-a to the medium. Cells imaged
before (Aa) and after (Ab) exposure to 10 l
M a1-PTH in a Ca

2+
-con-
taining medium. Note the marked increase in fluorescence intensity
in the cells’ cytosol, the large membrane blebs (arrows), and the
increase in projected area of the cells. Cells imaged before (Ac) and
after (Ad) exposure to 10 l
M a1-PTH in a Ca
2+
-free medium. Note
the absence of blebbing, but a similar increase in cells’ cytosol
fluorescence intensity. (B) Cells were imaged before (Ba) and after
(Bb) 50 l
M PIN-a exposure to a Ca
2+
-containing medium. In (Bc),
the bars indicate the relative projected area of the cells as a funct-
ion of PIN-a concentration and time of exposure. (C) Immunostain-
ing of actin under control conditions (Ca) and after exposure of the
cells to either 10 l
M a1-PTH (Cb) or 10 lM PIN-a. In (Cb), note the
distinct distribution and clumping of the immunolabelling.
P. Llanos et al. Toxic actions of puroindoline-a and a1-purothionin
FEBS Journal 273 (2006) 1710–1722 ª 2006 The Authors Journal compilation ª 2006 FEBS 1717
progressive recruitment of monomers would increase
the pore size. Under our conditions, once the cationic
pore is formed, the aggregation of further monomers
would not only augment the pore size, but also, by
exposing some amino-acid residues, create a new ani-
onic-selectivity filter. Although this may explain the
different channel behaviours detected with a1-PTH,

further experiments are needed to support this model.
In isolated mouse hemidiaphragm preparations,
a1-PTH completely blocked directly and indirectly
electrical muscle twitches. As a1-PTH caused mem-
brane depolarization, voltage-dependent sodium chan-
nels must be inactivated and unable to generate action
potentials in muscle fibres. a1-PTH effects can be rela-
ted to its predominant ability to form low-conductance
cationic-selective channels, as reported in liposomes.
The fact that increasing external Ca
2+
prevented mus-
cle depolarization by a1-PTH is not surprising because
pore formation induced by PTHs and PINs is inhibited
by high Ca
2+
concentrations [14,16]. Also, a1-PTH eli-
cited permeability changes that may depolarize nerve
terminals supplying the neuromuscular junction. Such
an action would explain the decrease in evoked trans-
mitter release (revealed by the decrease in EPP ampli-
tudes) and the increased spontaneous quantal release
(manifested by an enhanced MEPP frequency). Thus,
both presynaptic and postsynaptic permeability chan-
ges co-operate to block neuromuscular transmission
and muscle contraction evoked by nerve stimulation.
In contrast, no changes were detected with PIN-a,
except for hyperpolarization of the muscle membrane
when high concentrations were used. This is expected
if one considers that PIN-a increases membrane per-

meability mainly to K
+
ions and that the reversal
potential for K
+
ions is more negative than the resting
membrane potential of muscle fibres. Therefore, the
increased permeability induced by PIN-a will result in
K
+
outflux and, as a consequence, in membrane
hyperpolarization.
The consequences of the pore-forming ability of
a1-PTH and PIN-a were also evaluated in NG108-15
cells stained with the styryl dye FM1-43. This vital dye
partitions into the plasma membrane and does not
ordinarily ‘flip-flop’ across it [20]. During exposure of
cells to a1-PTH action, the fluorescent staining of the
cell’s membrane by the FM1-43 dye was particularly
useful for delineating membrane blebbing and follow-
ing the 3D projected area of the cells. The develop-
ment of blebs in the presence of a1-PTH is probably
related to the membrane permeability changes it indu-
ces, as, in other neuronal cells, blebbing has been asso-
ciated with raised intracellular Na
+
concentration [21].
Also, a1-PTH-treated NG108-15 cells exhibited an
increase in FM1-43 fluorescence intensity, similar to
that previously observed at the nodes of Ranvier of

myelinated axons [15]. This may be due to dye entry
into cells via a1-PTH-formed channels, as previously
reported for mechanotransducer channels [22]. Another
possibility is that the increased fluorescence intensity
reflects changes in membrane potential, as FM1-43 has
also been used as a voltage-sensitive dye [20]. In this
case, a 3.3% fluorescence increase per 100 mV poten-
tial change would be expected. This value is too low to
account for the marked increase in FM1-43 fluores-
cence intensity observed in NG108-15 cells. FM1-43
has also been reported to be a useful probe for monit-
oring phospholipid scrambling [23]. Taking into
account that one of the earliest detectable events in
cells undergoing apoptosis is phospholipid scrambling
[24], the FM1-43 fluorescence increase detected during
a1-PTH action, together with membrane blebbing and
disorganization and disarray of cytoskeletal actin, may
represent apoptotic events involved in a1-PTH cyto-
toxicity.
Experimental procedures
Extraction and purification of PIN-a and a1-PTH
PIN-a (M
r
12 920) and a1-PTH (M
r
4921.89) were puri-
fied from wheat seeds of Triticum aestivum sp., using a
modification of previously described procedures [18,25].
Briefly, 4 kg wheat endosperm flour was extracted with a
10-L solution containing 100 mm Tris buffer, 100 mm

NaCl, 5 mm EDTA and 5% Triton X-114 (pH 7.8). After
stirring (12 h, 4 °C) and centrifugation (8000 g, 30 min),
the supernatants were heated at 30 °C to allow phase par-
titioning, and the upper, detergent-poor phase was discar-
ded. The lower, detergent-rich phase was diluted with
5 vol. water and loaded on a column packed with a cation
exchanger (SP Biobeads; Pharmacia, Montigny-le-Breton-
neux, France). Proteins were eluted by applying a gradient
from 0.02 to 0.7 m NaCl in Tris buffer without Triton
X-114. Analysis of the collected fractions by SDS ⁄ PAGE
indicated that the PTHs were eluted as a single peak just
after the PINs. Separate pools of the PTH-containing and
PIN-containing crude fractions were dialyzed against de-
ionized water and freeze-dried, and a1-PTH, a2-PTH and
b-PTH were separated (at room temperature) by semipre-
parative RP-HPLC. The HPLC column was packed with
Nucleosil C18 (5 lm, 300 A
˚
), the PTHs were eluted with
an acetonitrile gradient (0.1% trifluoroacetic acid in deion-
ized water to 0.1% trifluoroacetic acid in acetonitrile), and
the fractions containing a1-PTH were pooled and freeze-
dried after dilution with deionized water. PIN-a was puri-
fied from the crude, freeze-dried, PIN-containing fraction
by cation-exchange chromatography on a 6 mL Resource
Toxic actions of puroindoline-a and a1-purothionin P. Llanos et al.
1718 FEBS Journal 273 (2006) 1710–1722 ª 2006 The Authors Journal compilation ª 2006 FEBS
S column (Pharmacia), as previously described [26]. The
homogeneity of the purified a1-PTH and PIN-a prepara-
tions was monitored by MS, as detailed by Elmorjani

et al. [27].
Reconstitution of PIN-a and a1-PTH into giant
liposomes
Giant liposomes were prepared by subjecting a mixture
(2 mL) of the protein (100 lg either PIN-a or a1-PTH) and
asolectin lipid vesicles (13 mm, in terms of lipid phospho-
rus) to a partial dehydration ⁄ rehydration cycle, as reported
previously [28]. After the partial dehydration ⁄ rehydration
cycle, the diameter of the resulting giant multilamellar lipo-
somes ranged from 5 to 100 lm.
Patch-clamp measurements
Aliquots (3–15 lL) of giant liposome preparations, in Petri
dishes (3.5 cm diameter), were mixed with 1 mL of the buf-
fer of choice (the bath solution) for electrical recording,
and unitary current recordings were performed using the
patch-clamp technique in an excised patch ‘inside out’ con-
figuration, as previously described [29]. Giga seals were
formed on giant liposomes with glass microelectrodes of 5–
10 MW resistance. After sealing, withdrawal of the pipette
from the liposome surface resulted in an excised patch.
Current was recorded with an EPC-9 patch-clamp amplifier
(Heka Elektronic, Lambrecht ⁄ Pfalzt, Germany) at a gain of
50–100 mVÆpA
)1
and a filter setting of 10 kHz. The holding
potential was applied to the interior of the patch pipette.
The bath potential was maintained at virtual ground via an
agar bridge (V ¼ V
bath
) V

pipette
), and the junction poten-
tial was compensated for when necessary. To study the
ionic selectivity of the protein-induced channels, we deter-
mined the relative ionic permeabilities from the reversal
potentials of the currents recorded in solutions of various
compositions, in response to potential ramps [ +150 to
)150 mV (60 mVÆs
)1
) or +100 to )100 mV (40 mVÆs
)1
)].
They were calculated from changes in reversal potentials,
brought about by ion replacement based on the Goldman–
Hodgkin–Katz flux equation [30,31]. The reversal potential
of a cationic current as a function of the concentration or
activity and the permeability of each ion species was calcu-
lated as previously described [32,33].
Patch-clamp data were analyzed off-line with TAC soft-
ware (Bruxton Corporation, Seattle, WA, USA) and Pulse
Fit (Heka Elektronic) software. All measurements were
made at  25 °C, and the pipette and bath solutions usu-
ally had the following composition:140 mm NaCl, 2.6 mm
CaCl
2
, 1.3 mm MgCl
2
, and 10 mm Hepes (adjusted to
pH 7.4 with NaOH). In some experiments, NaCl was
replaced by either KCl or sodium gluconate. All reagents

and chemicals were purchased from Sigma Biochemical Co.
(St Louis, MO, USA) and Merck (Darmstadt, Germany).
Electrophysiological and mechanical recordings
in isolated mouse hemidiaphragm muscles
Left and right hemidiaphragm muscles with their associated
phrenic nerves were isolated from adult Swiss-Webster mice
(20–25 g) killed by cervical vertebrae dislocation followed by
immediate exsanguination. Phrenic nerve ⁄ hemidiaphragm
muscle preparations were mounted in a RhodorsilÒ-lined
(Rhoˆ ne-Poulenc, St Fons, France) Plexiglas chamber (2-mL
or 4-mL capacity) and bathed in a standard physiological
solution gassed with pure O
2
and composed of 154 mm
NaCl, 5 mm KCl, 2 mm CaCl
2
,1mm MgCl
2
,11mm
glucose, and 5 mm Hepes (adjusted to pH 7.4 with NaOH).
Lyophilized PIN-a and a1-PTH were dissolved in 100 mm
Hepes buffer and stored as 1-mm stock solutions at )18 °C.
The stock solutions were diluted with the standard physiolo-
gical solution before experiments were performed at room
temperature. All experiments on mice were carried out in
accordance with the European Communities Council
Directive of 24 November 1986 (86/609/EEC), regarding the
ethical use of animals for experimental procedures.
Membrane potentials and synaptic potentials were recor-
ded with intracellular microelectrodes filled with 3 m KCl

(8–18 MW resistance), using conventional techniques and
an Axoclamp-2A system (Axon Instruments, Union City,
CA, USA). Recordings were made continuously from the
same endplate before and during treatment with the pro-
teins being tested. Electrical signals after amplification were
collected and digitized, at a sampling rate of 25 kHz, with
the aid of a computer equipped with an analogue-to-digital
interface board (DT 2821; Data Translation, Marlboro,
MA, USA). Computerized data acquisition and analysis
were performed with a program kindly provided by
J. Dempster (University of Strathclyde, Scotland, UK).
For twitch tension measurements, one of the tendons of
the hemidiaphragm muscle was tied with silk thread, via an
adjustable stainless steel hook, to an FT03 isometric transdu-
cer (Grass Instruments, West Warwick, RI, USA), and the
other tendon was pinned to the Rhodorsil-lined chamber.
Twitches were evoked either by stimulating the motor nerve
of isolated neuromuscular preparations via a suction micro-
electrode adapted to the diameter of the nerve, or by direct
muscle stimulation via an electrode assembly placed along
the length of the fibres. Pulses were supplied by a S-44 stimu-
lator (Grass Instruments) at frequencies of 0.2–40 Hz. For
each preparation investigated, the resting tension was
adjusted to obtain maximal contractile responses. Signals
from the isometric transducer were amplified, collected, and
digitized with the aid of a computer equipped with a DT
2821 analogue-to-digital interface board (Data Translation).
Cultured neuroblastoma cells
Rodent neuroblastoma (NG108-15) cells were grown in
monolayer cultures on glass coverslips using Dulbecco’s

P. Llanos et al. Toxic actions of puroindoline-a and a1-purothionin
FEBS Journal 273 (2006) 1710–1722 ª 2006 The Authors Journal compilation ª 2006 FEBS 1719
modified Eagle’s medium supplemented with 5% fetal
bovine serum, 100 lm hypoxanthine, 0.4 lm aminopterin,
16 lm thymidine, 2 mm glutamine, and 3 lm glycine, as des-
cribed previously [34]. Three days before the experiments,
cells were differentiated by adding 0.5 mm dibutyryl-cAMP
to the medium and reducing the serum concentration to
1%. Culture reagents were purchased from Sigma-Aldrich
Chimie (Saint Quentin-Fallavier, France) and Invitrogen
(Cergy Pontoise, France). The cultures were maintained at
37 °C in a humidified atmosphere containing 95% air ⁄
5% CO
2
.
Plasma membrane staining of neuroblastoma
cells
Before imaging, NG108-15 cells were rinsed free of the cul-
ture medium, and their plasma membrane was stained
(30 min, room temperature) with the styryl dye N-[3-(trieth-
ylammonium)propyl]-4-(4-dibutylaminostyryl pyridinium)
dibromide (FM1-43; Molecular Probes Europe Bv, Leiden,
the Netherlands; 2 lm) dissolved in standard physiological
solution. Thereafter, cells were rinsed with dye-free solu-
tion. The dry extracts of the two wheat endosperm proteins
were dissolved as described above, and added to the bath-
ing solution during the experiments.
Actin immunostaining of neuroblastoma cells
The culture medium of NG108-15 cells was added with
either PIN-a (10 lm)ora1-PTH (10 lm), and cells were

incubated for 2–4 h at 37 °C. Then, they were rinsed with
phosphate-buffered saline (NaCl ⁄ P
i
) and fixed with either
4% paraformaldehyde in NaCl ⁄ P
i
(15 min, 37 °C) or with
100% methanol (4 min, )20 °C). After being washed three
times with NaCl ⁄ P
i
to remove excess fixative, the cells were
permeabilized and blocked (l h, room temperature) with
NaCl ⁄ P
i
containing 0.1% Triton X-100 and 3% BSA
(blocking buffer). Subsequently, cells fixed by either method
were incubated with the primary JLA-20 antibodies (1 : 100
dilution; Jackson ImmunoResearch Laboratories, Inc, West
Grove, PA, USA), washed with NaCl ⁄ P
i
, and incubated
(1 h, room temperature) with Texas-red-conjugated secon-
dary antibodies (1 : 100 dilution; Molecular Probes). Cover-
slips were mounted on to glass slides with Vectashield
antifading mounting medium (Vector Laboratories, Inc,
Burlingame, CA, USA). In some experiments, AlexaFluor-
594-conjugated phalloidin (Molecular Probes) was used to
visualize F-actin in fixed cells.
Confocal laser scanning microscopy
Time-lapsed imaging of cells was performed using a Saras-

tro-2000 confocal system (Molecular Dynamics, Sunnyvale,
CA, USA) mounted on a Nikon Optiphot-2 upright
microscope. The system was controlled with imagespace
3.10 software and a Silicon Graphics Personal Iris
4D ⁄ 35G workstation (Mountain View, CA, USA). The
488 and 514 nm lines of an argon-ion laser were used for
fluorescence excitation of FM1-43 and Texas-red or Alexa-
Fluor-594, respectively. Images were collected at room
temperature using a water-immersion lens (· 40; numerical
aperture ¼ 0.75), and the aperture setting of the confocal
pinhole was 100 lm. A 3% neutral density transmission
filter was used in all experiments, and the photomultiplier
gain was kept constant during experiments. Series of
optical sections were collected using a standard scanning
mode format of 512 · 512 pixels, and ‘look through’ pro-
jections were constructed from the sections.
Statistical analysis
Values in the text are expressed as means ± SEM of n
experiments. Statistical analysis of data was performed
using Student’s t test (two-tailed). Data were considered
significant at P < 0.05.
Acknowledgements
This work was made possible by the ECOS Sud-CON-
ICYT (C03S02) exchange program, and was supported
in part by Fondecyt-Chile (grant No. 1040546), the
Centre National de la Recherche Scientifique and the
Institut National de la Recherche Agronomique. J.M.
was supported by the Fondation pour la Recherche
Me
´

dicale. We thank Dr M. Malo and Dr B. Rouzaire-
Dubois for providing the NG108-15 cells used in this
study, and Dr H. Rogniaux for performing MS.
Confocal microscopy studies were performed on the
Plate-forme Imagerie et Biologie Cellulaire of the
Gif-sur-Yvette Campus.
References
1 Douliez JP, Michon T, Elmorjani K & Marion D
(2000) Structure, biological and technological functions
of lipid transfer proteins and indolines, the major lipid
binding proteins from cereal kernels. J Cereal Sci 32,
1–20.
2 Tanchak MA, Schernthaner JP, Giband M & Altosaar I
(1998) Tryptophanins: isolation and molecular charac-
terisation of oat cDNA clones encoding proteins struc-
turally related to puroindoline and wheat grain softness
protein. Plant Sci 137, 173–184.
3 Gautier MF, Cosson P, Guirao A, Alary R & Joudrier
P (2000) Puroindoline genes are highly conserved in
diploid ancestor wheats and related species but absent
in tetraploid Triticum species. Plant Sci 153, 81–91.
Toxic actions of puroindoline-a and a1-purothionin P. Llanos et al.
1720 FEBS Journal 273 (2006) 1710–1722 ª 2006 The Authors Journal compilation ª 2006 FEBS
4 Florack DE & Stiekema WJ (1994) Thionins: properties,
possible biological roles and mechanisms of action.
Plant Mol Biol 26, 25–37.
5 Bohlmann H (1994) The role of thionins in plant pro-
tection. CRC Crit Rev Plant Sci 13, 1–16.
6 Le Bihan T, Blochet JE, Desormeaux A, Marion D &
Pezolet M (1996) Determination of the secondary struc-

ture and conformation of puroindolines by infrared and
Raman spectroscopy. Biochemistry 35, 12712–12722.
7 Blochet JE, Kaboulou A, Compoint JP & Marion D
(1991) Amphiphilic proteins from wheat flour: specific
extraction, structure and lipid binding properties. In
Gluten Proteins (Bushuk, W & Tkachuk, R, eds), pp.
314–325. American Association of Cereal Chemists,
St Paul, MN.
8 Dubreil L, Gaborit T, Bouchet B, Gallant DJ, Broeka-
ert WF, Quillien L & Marion D (1998) Spatial and
temporal distribution of the major isoforms of puroin-
dolines (puroindoline-a and puroindoline-b) and non
specific lipid transfer protein (ns-LTPe1) of Triticum
aestivum seeds. Relationships with their in vitro antifun-
gal properties. Plant Sci 138, 121–135.
9 Krishnamurthy K, Balconi C, Sherwood JE & Giroux
MJ (2001) Wheat puroindolines enhance fungal disease
resistance in transgenic rice. Mol Plant Microbe Interact
14, 1255–1260.
10 Kashimoto T, Sakakibara R, Huynh QK, Wada H &
Yoshizumi H (1979) The effect of purothionin on
bovine adrenal medullary cells. Res Commun Chem
Pathol Pharmacol 26, 221–224.
11 Carrasco L, Vazquez D, Hernandez-Lucas C, Carbon-
ero P & Garcia-Olmedo F (1981) Thionins: plant pep-
tides that modify membrane permeability in cultured
mammalian cells. Eur J Biochem 116, 185–189.
12 Kwak KB, Lee YS, Suh SW, Chung CS, Ha DB &
Chung CH (1989) Purothionin from wheat endosperm
reversibly blocks myogenic differentiation of chick

embryonic muscle cells in culture. Exp Cell Res 183,
501–507.
13 Oka T, Murata Y, Nakanishi T, Yoshizumi H, Hayash-
ida H, Ohtsuki Y, Toyoshima K & Hakura A (1992)
Similarity, in molecular structure and function, between
the plant toxin purothionin and the mammalian pore-
forming proteins. Mol Biol Evol 9, 707–715.
14 Hughes P, Dennis E, Whitecross M, Llewellyn D &
Gage P (2000) The cytotoxic plant protein, beta-puro-
thionin, forms ion channels in lipid membranes. J Biol
Chem 275, 823–827.
15 Mattei C, Elmorjani K, Molgo
´
J, Marion D & Benoit E
(1998) The wheat proteins puroindoline-a and alpha1-
purothionin induce nodal swelling in myelinated axons.
Neuroreport 9, 3803–3807.
16 Charnet P, Molle G, Marion D, Rousset M & Lullien-
Pellerin V (2003) Puroindolines form ion channels in
biological membranes. Biophys J 84, 2416–2426.
17 Llanos P, Henriquez M, Minic J, Elmorjani K, Marion
D, Riquelme G, Molgo
´
J & Benoit E (2004) Neuronal
and muscular alterations caused by two wheat endo-
sperm proteins, puroindoline-a and alpha1-purothionin,
are due to ion pore formation. Eur Biophys J 33, 283–
284.
18 Blochet JE, Chevalier C, Forest E, Pebay-Paryroula
E, Gautier MF, Pe

´
zolet M & Marion D (1993)
Complete amino acid sequence of puroindoline, a
new basic and cystine-rich protein with a unique
tryptophan-rich domain, isolated from wheat endo-
sperm by Triton X-114 phase partitioning. FEBS Lett
329, 336–340.
19 Shai Y (1999) Mechanism of the binding, insertion and
destabilization of phospholipid bilayer membranes by
alpha-helical antimicrobial and cell non-selective mem-
brane-lytic peptides. Biochim Biophys Acta 1462, 55–70.
20 Betz W, Mao F & Smith C (1996) Imaging exocytosis
and endocytosis. Curr Opin Neurobiol 6, 365–371.
21 Shi X, Gillespie PG & Nuttall AL (2005) Na
+
influx
triggers bleb formation on inner hair cells. Am J Physiol
Cell Physiol 288, C1332–C1341.
22 Gale JE, Marcotti W, Kennedy HJ, Kros CJ &
Richardson GP (2001) FM1-43 dye behaves as a per-
meant blocker of the hair-cell mechanotransducer chan-
nel. J Neurosci 15, 7013–7025.
23 Zweifach A (2000) FM1-43 reports plasma membrane
phospholipid scrambling in T-lymphocytes. Biochem J
349, 255–260.
24 Martin S, Reutelingsperger C, McGahon A, Rader J,
Van Schie R, LaFace D & Green D (1995) Early redis-
tribution of plasma membrane phosphatidylserine is a
general feature of apoptosis regardless of the initiating
stimulus: inhibition by overexpression of Bcl-2 and Abl.

J Exp Med 182, 1545–1556.
25 Dubreuil L, Compoint JP & Marion D (1997) Inter-
action of puroindolines with wheat flour polar lipids
determines their foaming properties. J Agric Food Chem
45, 108–116.
26 Biswas SC, Dubreil L & Marion D (2001) Interfacial
behaviour of wheat puroindolines: monolayers of pur-
oindolines at the air–water interface. Colloid Polymer
Sci 279, 607–614.
27 Elmorjani K, Lurquin V, Lelion A, Rogniaux H &
Marion D (2004) A bacterial expression system revisited
for the recombinant production of cystine-rich plant
lipid transfer proteins. Biochem Biophys Res Commun
316, 1202–1209.
28 Riquelme G, Lopez E, Garcia-Segura LM, Ferragut JA
& Gonzalez-Ros JM (1990) Giant liposomes: a model
system in which to obtain patch-clamp recordings of
ionic channels. Biochemistry 29, 11215–11222.
29 Hamill OP, Marty A, Neher E, Sakmann B & Sigworth
FJ (1981) Improved patch-clamp technique for high-
resolution current recording from cells and cell-free
P. Llanos et al. Toxic actions of puroindoline-a and a1-purothionin
FEBS Journal 273 (2006) 1710–1722 ª 2006 The Authors Journal compilation ª 2006 FEBS 1721
membrane patches. Pflu
¨
gers Arch Eur J Physiol 391,
85–100.
30 Goldman DE (1943) Potential, impedance, and rectifica-
tion in membranes. J General Physiol 27, 37–60.
31 Hodgkin AL & Katz B (1949) The effect of sodium ions

on the electrical activity of the giant axon of the squid.
J Physiol (London) 108, 37–77.
32 Chang DC (1983) Dependence of cellular potential on
ionic concentrations. Data supporting a modification
of the constant field equation. Biophys J 43, 149–
156.
33 Valera S, Hussy N, Evans RJ, Adami N, North RA,
Surprenant A & Buell G (1994) A new class of ligand-
gated ion channel defined by P2x receptor for extracel-
lular ATP. Nature 371, 516–519.
34 Rouzaire-Dubois B & Dubois JM (1990) Tamoxifen
blocks both proliferation and voltage-dependent K
+
channels of neuroblastoma cells. Cell Signal 2, 387–393.
Toxic actions of puroindoline-a and a1-purothionin P. Llanos et al.
1722 FEBS Journal 273 (2006) 1710–1722 ª 2006 The Authors Journal compilation ª 2006 FEBS