Tài liệu Báo cáo khoa học: Inhibitory effects of nontoxic protein volvatoxin A1 on pore-forming cardiotoxic protein volvatoxin A2 by interaction with amphipathic a-helix doc

Inhibitory effects of nontoxic protein volvatoxin A1
on pore-forming cardiotoxic protein volvatoxin A2
by interaction with amphipathic a-helix
Pei-Tzu Wu
1
, Su-Chang Lin
2
, Chyong-Ing Hsu
1
, Yen-Chywan Liaw
2
and Jung-Yaw Lin
1
1 Institute of Biochemistry and Molecular Biology, College of Medicine, National Taiwan University, Taipei, Taiwan
2 Institute of Molecular Biology Academia S inica, Taipei, Taiwan
Volvatoxin A (VVA) has been isolated from Volvari-
ella volvacea, and consists of volvatoxin A2 (VVA2)
and volvatoxin A1 (VVA1) [1]. VVA has several biolo-
gical activities, such as: (a) lysis of human red blood
cells; (b) swelling tumor cells and the mitochondria
of liver cells; (c) inhibition of protein biosynthesis;
and (d) causing cardiac arrest via activation of the
Ca
2+
-dependent ATPase enzyme in the ventricular
microsomal fraction [1–3]. The hemolytic activity of
VVA2 is totally inhibited by VVA1 at a molar ratio of
2 [4,5]. Previous studies have shown that VVA2 is a
b-pore-forming toxin, with a heparin-binding site
(HBS) encoded within the C-terminal b-strands (b6, b7
and b8). This HBS structure is indispensable for the

Keywords
amphipathic a-helix; co-pull-down
experiment; tandem repeat protein;
volvatoxin A1; volvatoxin A2
Correspondence
J Y. Lin, Institute of Biochemistry and
Molecular Biology, College of Medicine,
National Taiwan University, F9, no. 1,
Section 1, Jen-Ai Road, Taipei 10051,
Taiwan
Fax: +886 2 23415334
Tel: +886 2 23123456 (ext. 8206 ⁄ 8207)
E-mail:
Database
The nucleotide sequence reported in this
paper has been submitted to the
DDBJ ⁄ EMBL ⁄ GenBank databases under the
accession number AY952461
(Received 22 March 2006, revised 1 May
2006, accepted 17 May 2006)
doi:10.1111/j.1742-4658.2006.05325.x
Volvatoxin A2, a pore-forming cardiotoxic protein, was isolated from the
edible mushroom Volvariella volvacea. Previous studies have demonstrated
that volvatoxin A consists of volvatoxin A2 and volvatoxin A1, and the
hemolytic activity of volvatoxin A2 is completely abolished by volvatoxin
A1 at a volvatoxin A2 ⁄ volvatoxin A1 molar ratio of 2. In this study, we
investigated the molecular mechanism by which volvatoxin A1 inhibits the
cytotoxicity of volvatoxin A2. Volvatoxin A1 by itself was found to be
nontoxic, and furthermore, it inhibited the hemolytic and cytotoxic activit-
ies of volvatoxin A2 at molar ratios of 2 or lower. Interestingly, volvatoxin

A1 contains 393 amino acid residues that closely resemble a tandem repeat
of volvatoxin A2. Volvatoxin A1 contains two pairs of amphipathic a-heli-
ces but it lacks a heparin-binding site. This suggests that volvatoxin A1
may interact with volvatoxin A2 but not with the cell membrane. By using
confocal microscopy, it was demonstrated that volvatoxin A1 could not
bind to the cell membrane; however, volvatoxin A1 could inhibit binding
of volvatoxin A2 to the cell membrane at a molar ratio of 2. Via peptide
competition assay and in conjunction with pull-down and co-pull-down
experiments, we demonstrated that volvatoxin A1 and volvatoxin A2 may
form a complex. Our results suggest that this occurs via the interaction of
one molecule of volvatoxin A1, which contains two amphipathic a-helices,
with two molecules of volvatoxin A2, each of which contains one amphi-
pathic a-helix. Taken together, the results of this study reveal a novel
mechanism by which volvatoxin A1 regulates the cytotoxicity of volvatoxin
A2 via direct interaction, and potentially provide an exciting new strategy
for chemotherapy.
Abbreviations
FITC, fluorescein isothiocyanate; GSH, glutathione; GSP, gene-specific primer; HBS, heparin-binding site; RBC, red blood cell; VVA,
volvatoxin A; VVA1, volvatoxin A1; VVA2, volvatoxin A2; VVA1-CTD, volvatoxin A1 C-terminal domain (198–391 residues); VVA1-NTD,
volvatoxin A1 N-terminal domain (1–197 residues).
3160 FEBS Journal 273 (2006) 3160–3171 ª 2006 The Authors Journal compilation ª 2006 FEBS
membrane interaction of VVA2 [6]. Furthermore, the
VVA2-binding receptor on the cell membrane has been
shown to be a sulfated glycosaminoglycan, as demon-
strated by affinity column chromatography [6]. Binding
of VVA2 to the cell membrane induced a protein con-
formational change of the VVA2 amphipathic a-helices
to form a prepore complex [7–10]. Therefore, the
amphipathic a-helices play an important role in VVA2
oligomerization and pore formation [6].

Pore-forming toxins are essentially naturally occur-
ring biological weapons produced by both prokaryotes
and eukaryotes, and include well-known toxins such as
diphtheria and anthrax toxins, as well as the less well-
known a-hemolysin and equinatoxin II [11–16]. An
important objective is to provide an effective inhibitor
of these virulence factors, and naturally occurring sub-
stances represent one potential source. Recently, there
has been much interest in the potential application of
these toxins to chemotherapy and the delivery of drugs
[17,18].
In an attempt to determine the inhibitory mechan-
ism of VVA1 on VVA2, we deduced the amino acid
sequence of VVA1 from the cDNA nucleotide
sequence. The primary structure of VVA1 is similar to
that of a tandem repeat form of VVA2, and the pre-
dicted secondary structure showed that it contains
two pairs of amphipathic a-helices. VVA1 completely
inhibited the hemolytic and cytotoxic activities of
VVA2 at VVA2 ⁄ VVA1 molar ratios of 2 or lower.
Taken together, our results provide evidence that
VVA1 interacts with VVA2 and regulates the cytotoxic
pore-forming activity of VVA2.
Results and Discussion
Characteristics of VVA1 structure
To study the structure of VVA1, we cloned VVA1
cDNA by the RACE method, as described previously
(supplementary Table S1) [19]. The coding region of
the cloned VVA1 cDNA contained 1179 nucleotides,
and the deduced amino acid sequence was identical to

that determined previously by protein sequencing (sup-
plementary Fig. S1) [5]. Interestingly, the amino acid
sequence of VVA1 was very similar to that of a tan-
dem repeat of VVA2. The N-terminal half fragment of
VVA1, designated volvatoxin A1 N-terminal domain
(VVA1-NTD) (1–197 residues), had 46.3% similarity
with VVA2 (Fig. 1A), whereas the C-terminal frag-
ment, designated volvatoxin A1 C-terminal domain
(VVA1-CTD) (198–391 residues), displayed 49.2%
similarity to VVA2 (Fig. 1A). The similarity between
VVA1-NTD and VVA1-CTD is 42.6%. The tertiary
structure of VVA2 shows that it has a pair of amphi-
pathic a-helices, denoted a-helix-C and a-helix-D [24],
which are essential for VVA2 dimerization [6]. Interest-
ingly, VVA1 also contains a pair of amphipathic
a-helices similar to VVA2 (Fig. 1A) (supplementary
Fig. S2). It has been shown that the amphiphilicity of
the amphipathic a-helix of VVA2 is indispensable for
protein interaction and oligomerization [6]. Secondary
structure analysis of VVA1 suggests that there might
be two pairs of amphipathic a-helices in both the
N-terminal and C-terminal domains. The hydrophobic
moments of amphipathic a-helix-C and a-helix-D of
VVA1-NTD were calculated to be 0.4 and 0.57,
respectively, while those of amphipathic a-helix-D¢ and
a-helix-E¢ of VVA1-CTD were 0.49 and 0.57 [25].
VVA2 has a basic HBS at its C-terminus that is
located within its b-strand, is indispensable for binding
to cell membranes, and has a pI value of 9.6, similar
to that of the snake venom cardiotoxin [6,20–22]. Nei-

ther VVA1-NTD nor VVA1-CTD has a basic HBS at
their C-terminus as VVA2 does. Additionally, the pI
values of the corresponding C-terminal regions of
VVA1-NTD and VVA1-CTD were found to be 4.3
and 4.6, respectively, suggesting that VVA1 has very
weak, if any, affinity for the anionic surface of cell
membranes [23]. Furthermore, we demonstrate here
that VVA1 has no noticeable affinity for simple lipid
membranes (Fig. 1B). When VVA1 was incubated with
liposomes and then centrifuged and electrophoresed,
analysis of the supernatant and pellet fractions showed
that VVA1 cannot bind to these simple membranes
(Fig. 1B, lanes 3 and 4). Additionally, VVA2 binds
liposome-containing membranes, as demonstrated by
the exclusive presence of VVA2 in the pellet fraction
(Fig. 1B, lanes 5 and 6). Intriguingly, the presence of
VVA1 inhibited the oligomerization of VVA2 and thus
its binding to simple lipid membranes (Fig. 1B, lanes 1
and 2). Therefore, this investigation of the structural
and binding characteristics of VVA1 indicates that
VVA1 may have the capacity for protein–protein inter-
action with VVA2 via its amphipathic a-helices, but it
is unlikely that VVA1 would be able to bind to the
membrane of cells.
Inhibitory effects of VVA1 on the hemolytic and
cytotoxic activity of VVA2
The effects of VVA1 on the hemolytic activity of
VVA2 were studied by incubation of human red blood
cells (RBCs) with the purified proteins. VVA1 itself
had no hemolytic activity when incubated with human

RBCs (Fig. 2A, column 2). Strikingly, VVA1 com-
pletely abolished the hemolytic activity of VVA2 at
P T. Wu et al. VVA1 is a novel toxin regulator of VVA2
FEBS Journal 273 (2006) 3160–3171 ª 2006 The Authors Journal compilation ª 2006 FEBS 3161
A
B
liposomes
VVA2
+
+–
–
–
––
–
–
+
S
12345678
PSP SP
+
++
++
+
VVA1
(MkDa)
(4µg)
(4µg)
200
116
97

VVA1
VVA2
S:supernatant
P:pellet
(monomer)
66
45
31
21
Fig. 1. Characteristics of volvatoxin A1 (VVA1) structure. (A) Alignment of the deduced amino acid sequence of VVA1 N-terminal domain
(VVA1-NTD) and VVA1 C-terminal domain (VVA1-CTD) with that of VVA2 (GenBank accession number AY362729). The secondary structural
elements of VVA1-NTD predicted by the PROF
SEC program are illustrated at the top of the sequence (orange), and those of VVA2 (green)
from the X-ray crystallographic analysis are shown below; the arrows represent b-strands, and the rods represent a-helices. Secondary struc-
tural elements of VVA1-NTD with PROF scores below 5 are shown in light orange. The completely conserved residues are shaded in dark
green, and similar aligned residues are shaded in pink. The residue numbers are indicated on the right. The ‘+’ symbol represents the amino
acid residues of the heparin-binding site (HBS) of VVA2 (166–194). (B) Inhibitory effects of VVA1 on the binding of VVA2 to liposomes. After
incubation of VVA1, VVA2 or the mixture of VVA2 and VVA1 at a molar ratio of 2 with liposomes (5 m
M)at37°C for 30 min, the reaction
mixtures were subjected to centrifugation at 100 000 g at 4 ° C for 1 h. The presence of VVA1 or VVA2 in the supernatant and pellet were
analyzed by 10% SDS ⁄ PAGE and visualized by Coomassie Blue staining.
VVA1 is a novel toxin regulator of VVA2 P T. Wu et al.
3162 FEBS Journal 273 (2006) 3160–3171 ª 2006 The Authors Journal compilation ª 2006 FEBS
VVA2 ⁄ VVA1 molar ratios of 2 or lower (Fig. 2A, col-
umns 3–5), while at a molar ratio of 4 the hemolytic
activity of VVA2 was reactivated (Fig. 2A, column 6).
To examine whether VVA1 affects the cytotoxicity
of VVA2, HeLa cells were treated with VVA2 (17 nm,
IC
50

of VVA2) and various amounts of VVA1 at
37° C for 24 h. Similar to the results obtained in
hemolytic experiments, VVA1 itself had no cytotoxicity
(Fig. 2B, column 2), but was able to inhibit the cyto-
toxicity of VVA2 completely at a VVA2 ⁄ VVA1 molar
ratios of 2 or lower (Fig. 2B, columns 3–5). Further-
more, the cytotoxicity of VVA2 was reactivated at a
molar ratio of 4 when it was incubated with HeLa cells
(Fig. 2B, column 6).
Additionally, confocal microscopy was employed to
study the inhibitory effects of VVA1 on VVA2. The
results showed that VVA1 by itself was unable to bind
to cell membranes (Fig. 3, panel FITC-VVA1 and pan-
els a–d). Moreover, preincubation of VVA2 and VVA1
(at a molar ratio of 2) inhibited VVA2 binding to the
cell membrane (Fig. 3, panels e–h). These results
strongly suggest that VVA1 inhibits the cytotoxicity of
VVA2 by preventing the binding of VVA2 to the cell
membrane.
Interactions between VVA1 and VVA2
To find whether direct interaction between VVA1 and
VVA2 is required for the inhibitory effects of VVA1 on
the pore-forming activity of VVA2, pull-down experi-
ments were performed. As a preliminary experiment, we
investigated the effects of different buffer constituents
on the oligomerization of VVA2. Only Triton X-100,
and not deoxycholate as had been reported for Bcl-2
family members, was able to induce the oligomerization
of VVA2 (supplementary Fig. S3A) [26]. Interestingly,
not even the harsh, denaturing environment of electro-

phoresis through the SDS ⁄ PAGE system could affect
VVA2 oligomer formation. Furthermore, the Triton
X-100 induction of oligomerization of VVA2 was able
to mimic the amphipathic environment of an artificial
cell membrane of liposomes (supplementary Fig. S3B)
[26–29]. Additionally, incubation of VVA2 without lipo-
somes inhibited their oligomerization (supplementary
Fig. S3B, lane 4). Thus the environment of detergent
micelles set up by buffering with Triton X-100 was very
similar to the natural environment and was used for
further experiments to determine interactions between
VVA2 and VVA1.
The input controls for the representative pull-down
experiment shown in Fig. 4A are lanes 11 and 12,
where the oligomerization of VVA2 in lane 11 is
clearly shown. Intriguingly, when VVA2 and VVA1 at
a molar ratio of 2 : 1 were preincubated together and
then electrophoresed, no oligomerization of VVA2 was
evident (Fig. 4A, lane 12). This result seemed to sug-
gest that there was indeed some form of interaction
between VVA1 and VVA2.
Furthermore, when beads linked to VVA2 were
incubated with VVA1 and then washed, eluted and
run on a 10% SDS ⁄ PAGE gel, VVA1 had clearly
bound to VVA2 (Fig. 4A, lane 1). Furthermore, when
a 2 : 1 molar mixture of VVA2 and VVA1 was incuba-
ted with VVA2-linked beads, both proteins were
adsorbed, and after elution both proteins were detected
A
B

Fig. 2. Effects of volvatoxin A1 (VVA1) on the hemolytic and cyto-
toxic activity of volvatoxin A2 (VVA2). (A) The hemolytic activity of
VVA2 regulated by VVA1. VVA2 (45 n
M) and various concentrations
of VVA1 were preincubated as indicated, and the percentage of
hemolysis was calculated as described in Experimental procedures.
Each value represents the mean ± SD of three independent experi-
ments. (B) VVA2 cytotoxicity was affected by VVA1. HeLa cells
were treated with VVA2 (17 n
M) and various amounts of VVA1 at
37 °C for 24 h. Cell death was assayed by using a trypan blue
exclusion assay [41]. Means ± SD are shown for three independent
experiments.
P T. Wu et al. VVA1 is a novel toxin regulator of VVA2
FEBS Journal 273 (2006) 3160–3171 ª 2006 The Authors Journal compilation ª 2006 FEBS 3163
via SDS ⁄ PAGE analysis (Fig. 4A, lane 2). Interest-
ingly, again no oligomers of VVA2 were detected after
incubation of VVA2 with VVA1 at a molar ratio of 2
(Fig. 4A, lane 2).
When VVA1 beads were incubated with VVA2,
VVA2 oligomers were adsorbed and eluted (Fig. 4A,
lane 4). Additionally, when VVA1 beads were incuba-
ted with the mixture of VVA2 and VVA1 (molar ratio
2 : 1), VVA2 and VVA1 were detected, but again, no
VVA2 oligomer was found (Fig. 4A, lane 5). Taken
together, these results strongly support the notion that
there is a direct interaction between VVA1 and VVA2,
and that at a VVA2 ⁄ VVA1 molar ratio of 2, VVA1 is
able to inhibit the oligomerization of VVA2. Extending
these results, we hypothesized that the inhibition of

VVA2 cytotoxic pore formation by VVA1 can only
occur at the ideal ratio of 2 : 1 or lower, due to the
ability of one molecule of VVA1 to interact with two
molecules of VVA2. At a higher ratio of VVA2 to
VVA1, the latter is not able to prevent VVA2 oligo-
merization and thus cannot inhibit VVA2 cytotoxicity.
Number of VVA1-binding sites for VVA2
To further identify the binding characteristics and to
investigate the dynamic interaction between VVA1 and
VVA2, we carried out co-pull-down experiments. The
amphipathic a-helix of VVA2 had previously been
a
b
c d
h
g
f
e
Fig. 3. Volvatoxin A1 (VVA1) inhibited volvatoxin A2 (VVA2) binding to the cell membrane. Binding of VVA2 to the cell membrane was abol-
ished by the presence of VVA1. HeLa cells were treated with fluorescein isothiocyanate (FITC)–VVA1, VVA2 and both at a molar ratio of 2,
as described in Experimental procedures. VVA1 was conjugated with FITC (green fluorescence, panel FITC–VVA1), and VVA2 was stained
with Alexa-568 (red fluorescence, panels a and e), while the nucleus was stained with Hoechst 33258 (blue color, panels b and f). The over-
lay of both images is shown in panels c and g. The phase-contrast image shows cellular morphology (phase panel). Bar, 40 lm.
VVA1 is a novel toxin regulator of VVA2 P T. Wu et al.
3164 FEBS Journal 273 (2006) 3160–3171 ª 2006 The Authors Journal compilation ª 2006 FEBS
A
B
Fig. 4. Interaction between volvatoxin A1 (VVA1) and volvatoxin A2 (VVA2) in the presence of Triton X-100. (A) Pull-down experiments.
VVA1 (45 n
M), VVA2 (45 nM) or the mixture (VVA2, 45 nM, and VVA1, 22.5 nM) were incubated with VVA2 beads, VVA1 beads or BSA beads

at 37 °C for 30 min in 0.02% Triton X-100. The beads were washed, and the bound proteins were eluted. The protein eluents were identi-
fied by 10% SDS ⁄ PAGE and visualized by silver staining. (B) Co-pull-down experiments. Linear oligomeric VVA2 (VVA1 beads) was prepared
from VVA1 beads, which were treated with VVA2 in 0.02% Triton X-100 buffer, and VVA1 (VVA2 beads) was prepared from VVA2 beads,
which were treated with VVA1 in the same buffer as described in Experimental procedures. The linear oligomeric VVA2 (VVA1 beads) was
then incubated with various amounts of VVA1, while the VVA1 (VVA2 beads) was incubated with various amounts of VVA2. The reaction
products were eluted with 0.5% SDS loading buffer, and the proteins in the eluents were analyzed by 10% SDS ⁄ PAGE and visualized by
silver staining.
P T. Wu et al. VVA1 is a novel toxin regulator of VVA2
FEBS Journal 273 (2006) 3160–3171 ª 2006 The Authors Journal compilation ª 2006 FEBS 3165
identified as necessary for the oligomerization of
VVA2 [6]. Furthermore, we had determined that
VVA1 encoded two regions that displayed a reason-
ably high degree of conservation to the VVA2 olig-
omerization domain. Thus, we hypothesized that
VVA1 may contain two binding sites for complex
formation with VVA2. To address this issue, co-pull-
down experiments were carried out. First, VVA1-
linked beads (VVA1 beads) were incubated with 45 nm
VVA2 [referred to as linear oligomeric VVA2 (VVA1
beads)]. This mixture was then incubated with various
amounts of VVA1 in the presence of 0.02% Triton
X-100, and eluted with 0.5% SDS buffer (Fig. 4B).
The results demonstrated that no VVA1 could be
bound to a VVA1 bead that was saturated with VVA2
oligomers (Fig. 4B, lanes 1–4), which may imply that
one molecule of VVA2 has one binding site for
interacting with either VVA1 or VVA2. Additional
investigation of the characteristics of binding of VVA2
to VVA1 will be necessary to further understand this
important interaction.

Next, we utilized VVA2-linked beads (VVA2 beads)
and incubated them with 4 nm VVA1 protein [referred
to as VVA1 (VVA2) beads] (Fig. 4B). VVA1 (VVA2
beads) was incubated with increasing amounts of
VVA2, and visualization on an SDS ⁄ PAGE gel showed
that the adsorbed VVA2 had oligomerized. As the
amount of VVA2 in the reaction was increased, more
VVA2 was bound to the VVA1 (VVA2 beads) in the
form of oligomers (Fig. 4B, lanes 7–10). Collectively,
these data indicate that one molecule of VVA1 has two
binding sites for interaction with two molecules of
VVA2, and that large amounts of free VVA2 can use
VVA1 as a basis for the formation of VVA2 oligomers.
Interaction of VVA1 and VVA2 by amphipathic
a-helix
To identify the binding sites in VVA1 responsible for
direct interaction with VVA2, peptide competition
assays, pull-down experiments and western blots were
carried out. For the peptide competition assays, the
amphipathic a-helices of VVA1 were generated as
recombinant peptide fragments denoted as reVVA1-
NTD-aH-C-D (amino acids 72–109) and reVVA1-
CTD-aH-D¢-E¢ (amino acids 260–302) (Fig. 5A). These
fragments were then used to compete with bead-linked
VVA1 for binding to VVA2. The effectiveness of com-
petition was interpreted via the amount of VVA2 bind-
ing to the bead-linked VVA1 after pull-down and
SDS ⁄ PAGE electrophoresis. The results showed that
the interaction of VVA2 with VVA1 beads was subject
to competition by the N-terminal helix pair (reVVA1-

NTD-aH-C-D) at a reVVA1-NTD-aH-C-D ⁄ VVA2
molar ratio of 10 (Fig. 5B, lanes 1–3). Interestingly,
the peptide fragment containing the C-terminal helix
pair (reVVA1-CTD-aH-D¢-E¢) was able to efficiently
compete with binding of VVA2 to the bead-linked
VVA1 at a reVVA1-CTD-aH-D¢-E¢⁄VVA2 molar ratio
of 2.5 (Fig. 5B, lanes 4–6). Furthermore, the reHBSF
peptide fragment could not compete with the interac-
tion of VVA2 with VVA1 beads (Fig. 5B, lanes 7–9).
This was an expected result, as the HBS fragment in
VVA2 was identified as the membrane-binding domain
[6]. These results suggest that the N-terminal and the
C-terminal pair of a-helices of VVA1 can bind to
VVA2 independently of each other and thus enable the
direct binding by one molecule of VVA1 of two mole-
cules of VVA2. This further complements our previous
results suggesting an optimal molar ratio of 2 for bind-
ing of VVA2 to VVA1. The anti-VVA2 IgG used in
this experiment only detects VVA2 and does not cross-
react with VVA1 (supplementary Fig. S4).
In the present study, we have shown that VVA1 com-
pletely inhibits the biological activity of VVA2 in vitro
at VVA2 ⁄ VVA1 molar ratios 2 or lower. This begs the
question of why a mushroom would produce a toxin
and at the same time an antidote. We believe that the
major reason why Volvariella volvacea produces VVA1
is so that it can associate with and, at the right ratio,
enhance the toxicity of VVA2. As shown previously, the
LD
50

of VVA1 or VVA2 individually is higher than
40 mgÆ(kg body weight)
)1
. At a molar ratio of 2, the
LD
50
of VVA2 ⁄ VVA1 is reduced to 6 mgÆ(kg body
weight)
)1
. However, at a molar ratio of 6, which is
similar to that in the mushroom, a still lower LD
50
was
evident. This intriguing phenomenon requires further
investigation [1].
On the basis of the present findings, we propose that
one molecule of VVA1 interacts with two molecules of
VVA2 and thus inhibits the formation of the mature
pore complex. Furthermore, we suggest that manipula-
tion of the levels of VVA1 may be utilized to inhibit
VVA2 oligomerization and pore formation until cer-
tain conditions are present to make it biologically
valuable. For example, it has recently been proposed
that native or recombinant pore-forming toxin may be
used as a biotherapeutic agent [30–32].
The novel approach of using pore-forming toxins for
the treatment of solid tumors, which have proven to be
quite resistant to conventional toxins [33–35], shows
great promise. One of the major drawbacks of using
these toxins is that they must be able to preserve the

main characteristics of the toxin during the transport
process in vivo [32,36]. Therefore, a targeted VVA1–
VVA2 complex may be introduced to the host as a
VVA1 is a novel toxin regulator of VVA2 P T. Wu et al.
3166 FEBS Journal 273 (2006) 3160–3171 ª 2006 The Authors Journal compilation ª 2006 FEBS
protoxin, thus having no toxicity to the animal, but with
the ability to target a tumor. Once at the appropriate
site, the VVA1 molecule could be dissociated, allowing
the VVA2 molecules to oligomerize and reactivate their
cytotoxic pore-forming activity. This has been demon-
strated previously, when a mutated anthrax protoxin
was cleaved by urokinase plasminogen activator and
selectively killed a subset of cancer cells that highly
expressed plasminogen activator [31,37–39]. Thus, this
description of a naturally occurring inhibitor of VVA2
represents a significant discovery, although its import-
ance in a clinical setting remains to be investigated.
Experimental procedures
Materials
Taq DNA polymerase and the pGEM-T vector were
obtained from Promega (Madison, WI). Restriction endo-
nucleases and T4 DNA ligase were from New England
Biolabs Inc. (Beverly, MA). The Marathon cDNA amplifi-
cation kit was from Clontech (Palo Alto, CA). Fluorescent
Alexa-568-labeled goat anti-rabbit and fluorescein isothio-
cyanate (FITC) were purchased from Chemicon Inter-
national (Temecula, CA). CNBr-activated Sepharose 4B,
glutathione (GSH)-agarose-4B column and pGEX-2T vec-
tor were from Amersham Biosciences (Uppsala, Sweden).
All other chemicals were of analytical grade.

Purification, and cDNA cloning of VVA1
VVA1 was purified from mushroom, V. volvacea, and the
amino acid sequence of VVA1 was determined by protein
techniques as reported previously [1,19]. The peptides
of VVA1 generated by N-tosyl-l-phenylalanine chloro-
methylketone treated-trypsin, Streptococcus aureus V8
endoproteinase or Lys-C endoproteinase digestion were
fractionated by HPLC with a C18 reverse-phase column
A
B
Fig. 5. Peptide competition assay. (A) Schematic representation of peptide competitors. (B) Binding of volvatoxin A2 (VVA2) to volvatoxin A1
(VVA1) was inhibited by the amphipathic a-helices of VVA1. The VVA2 and VVA1 mixture (molar ratio 2) was incubated with VVA1 beads;
the interaction was examined in the presence of increasing amounts of competitors. The adsorbed protein was analyzed by western blots
using anti-VVA2 IgG; this indicated that the two amphipathic a-helices competed for the interaction between VVA2 and VVA1 beads.
P T. Wu et al. VVA1 is a novel toxin regulator of VVA2
FEBS Journal 273 (2006) 3160–3171 ª 2006 The Authors Journal compilation ª 2006 FEBS 3167
(4.6 · 250 mm) that was eluted with a linear gradient of
acetonitrile (0–80%) in 0.1% trifluoroacetic acid. The
polypeptides obtained by HPLC were subjected to amino
acid sequence analysis using an ABI 476 A Applied
Biosystems (Foster City, CA) automated amino acid
sequencer [19]. The amino acid sequence of VVA1 was
used for the design of degenerate primers and cloning the
cDNA of VVA1. All primers used in this study are repor-
ted in supplementary Table S1.
Poly(A
+
) RNA was isolated from the total RNA frac-
tion on an oligo (dT)-cellulose column, and poly(A
+

)-rich
mRNA was reverse-transcribed with a Marathon cDNA
amplification kit [19,40]. The cDNAs were ligated to
Marathon adaptors for 5¢ and 3¢ rapid amplification of
cDNA ends (RACE), and the products were used as the
template for subsequent PCR. In the first PCR, VVA1
cDNA was amplified with the sense degenerate primer A
and the antisense degenerate primer B, corresponding to
amino acid residues 1–6 and 385–391 of VVA1, respect-
ively. The amplified first PCR products were used as tem-
plate for nested PCR with the sense degenerate primer A
and the antisense degenerate primer C, which corresponds
to amino acid residues 163–168 of VVA1. The products
of this second PCR were sequenced and used to design
the specific antisense primers GSP-1, corresponding to
amino acid residues 40–47 of VVA1, and GSP-3, corres-
ponding to amino acid residues 24–31 of VVA1. In the
third PCR, GSP-1 was used along with the Marathon
primer AP-1, and the products were used as the template
for the fourth PCR, in which GSP-3 was used along with
the sense primer AP-2 to obtain the 5¢-end of the VVA1
cDNA.
The products of the second PCR were used to design
GSP-2 and GSP-4 specific sense primers corresponding to
amino acid residues 128–136 and 155–162 of VVA1,
respectively. The two primers were used along with the
Marathon primers AP-1 and AP-2 to yield the 3¢-end of the
VVA1 cDNA.
The full-length VVA1 cDNA was obtained by amplifying
V. volvacea cDNAs with the sense primer GSP-5, which

encodes the start codon and the first eight N-terminal
amino acid residues of VVA1, and the antisense primer
GSP-6, which encodes the last eight C-terminal amino acid
residues and the stop codon of VVA1 (GenBank accession
number AY952461). The PCR products were then ligated
into the T vector.
Liposome-binding assay
Liposomes were prepared as described previously [6].
After incubation of VVA1 (4 lg), VVA2 (4 lg) or the
mixture of VVA2 and VVA1 (at a molar ratio 2) with
liposomes (5 mm)at37°C for 30 min, the reaction mix-
tures were subjected to centrifugation at 100 000 g at
4 °C for 1 h (Beckman TLA 100.2; Beckman Coulter,
Taipei, Taiwan). Then, the supernatant and pellet were
analyzed by 10% SDS ⁄ PAGE and visualized by Coomas-
sie Blue staining.
Hemolytic activity assay
Human RBCs were prepared by washing three times with
NaCl ⁄ P
i
(137 mm NaCl, 1.5 mm KH
2
PO
4
, 2.7 mm KCl,
8.1 mm Na
2
HPO
4
, pH 7.4) [6]. VVA2 (45 nm, inducing

50% hemolysis) and various amounts of VVA1 were pre-
mixed in NaCl ⁄ P
i
, and then 0.1 mL of human RBCs
(3 · 10
7
cellsÆmL
)1
) was added. The reaction mixtures were
further incubated at 37 °C for 30 min, and the reaction
was terminated by centrifuging at 13 000 g for 5 min
(KUBOTA RA-155; Kubota, Osaka, Japan). The superna-
tant was measured at 540 nm to determine the degree of
hemolysis. One hundred per cent hemolysis was defined as
the same volume of the human red blood cells in the pres-
ence of 0.1% Triton X-100 [6].
Cytotoxicity assay
HeLa cells were grown in DMEM supplemented with 10%
FBS, 2 mml-glutamine, 100 unitsÆmL
)1
penicillin, and
100 lgÆmL
)1
streptomycin (Life Technologies, Inc.) under
5% CO
2
at 37 °C. HeLa cells (3 · 10
5
) were then treated
with mixtures of VVA2 (17 nm, causing 50% cytotoxicity)

and various amounts of VVA1 for 24 h. The cells were then
trypsinized, collected, and treated with 2% trypan blue dye
in NaCl ⁄ P
i
at 37 °C for 5 min. The surviving cells were
counted with a hemocytometer [41,42].
Confocal microscope analysis
FITC–VVA1 was prepared by coupling 1.5 mgÆmL
)1
VVA1
with 40 lgÆmL
)1
FITC in 0.1 m NaHCO
3
at 4 °C for 16 h.
The free FITC was removed with a Sephadex G25 column,
and the effect of FITC–VVA1 on the hemolytic activity of
VVA2 was shown to be the same as that of VVA1. HeLa
Cells (4 · 10
5
) grown on coverslips were treated with FITC-
conjugated VVA1 (17 nm), VVA2 (17 nm) or the mixture of
VVA2 (17 nm) and VVA1 (8.5 nm). Immunostaining was
performed by fixing the cells with 4% paraformaldehyde in
1 · NaCl ⁄ P
i
on coverslips. To detect VVA2 protein, nonper-
meabilized fixed cells were blocked in blocking buffer (10%
FBS in NaCl ⁄ P
i

) for 30 min. The cells were then probed with
anti-VVA2 (1 : 1000) at room temperature for 1 h. After
extensive washing in NaCl ⁄ P
i
, the washed cells were stained
with Alexa-568-conjugated goat anti-rabbit IgG (1 : 1000)
for 60 min. During the last 15 min of secondary antibody
staining, Hoechst 33258 (5 lgÆmL
)1
) was applied for observa-
tion of the nucleus. After washing with NaCl ⁄ P
i
, slides were
mounted in mounting solution (80% glycerol in NaCl ⁄ P
i
),
and sealed with nail polish. The cells were subjected to
VVA1 is a novel toxin regulator of VVA2 P T. Wu et al.
3168 FEBS Journal 273 (2006) 3160–3171 ª 2006 The Authors Journal compilation ª 2006 FEBS
immunostaining for observation of the VVA1 and VVA2 as
described previously [41].
Pull-down experiment
VVA2, VVA1 or BSA (8 mg) was coupled to CNBr-Seph-
arose beads (Amersham Pharmacia Biotech) in coupling
buffer (100 mm NaHCO
3
, pH 8.3, and 500 mm NaCl) and
incubated at 4 °C overnight. Residual active groups were
blocked with 1 m ethanolamine, pH 8.0, at room tempera-
ture for 2 h. The beads were then washed four times with

alternating pH buffers. Each wash cycle consisted of Tris
buffer (0.1 m Tris, pH 8.0, and 500 mm NaCl) and acid
wash buffer (0.1 m sodium acetate, pH 4.0, and 500 mm
NaCl) [43]. Thirty microliters of protein-conjugated beads
was incubated with 45 nm VVA1, VVA2 or a mixture of
VVA2, 45 nm, and VVA1, 22.5 nm,at37°C for 30 min in
0.02% Triton X-100, 50 mm Tris, pH 8.0. The beads were
washed three times with 50 mm Tris buffer and eluted with
0.5% SDS loading buffer (50 mm Tris, pH 8.0, 0.5% SDS,
10% glycerol and 0.03% bromophenol blue). The eluent
was analyzed by 10% SDS ⁄ PAGE and visualized by silver
staining.
Co-pull-down experiment
For further study of the protein–protein interaction, we
performed a co-pull-down experiment to analyze the num-
ber of binding sites of VVA1 and VVA2. VVA1 (VVA2
beads) was prepared by incubating VVA2 beads with 45 nm
VVA1 at 37 °C for 30 min. Adsorbed proteins were ana-
lyzed as mentioned above. After removal of the unbound
VVA1, the VVA1 (VVA2 beads) was incubated with var-
ious amounts of VVA2 in 0.02% Triton X-100 at 37 °C for
30 min. Linear oligomeric VVA2 (VVA1 beads) was pre-
pared by incubating VVA1 beads with 45 nm VVA2 at
37 °C for 30 min. After removal of the unbound VVA2 by
washing with 50 mm Tris buffer, the linear oligomeric
VVA2 (VVA1 beads) was incubated with various amounts
of VVA1 in 0.02% Triton X-100 at 37 °C for 30 min. The
reaction was terminated by centrifugation, and the beads
were washed three times with 50 mm Tris buffer, and eluted
with 0.5% SDS loading buffer. The eluents were subjected

to 10% SDS ⁄ PAGE analysis.
Peptide competition assay
To determine the protein–protein interaction site for VVA1
and VVA2, the amphipathic a-helix regions of VVA1-NTD
(72–109 residues of VVA1) and VVA1-CTD (260–302 resi-
dues of VVA1) were PCR amplified (supplementary Table
S1), and then ligated into the pGEX-2T vector for protein
expression. The HBS fragment of VVA2 (165–199 residues
of VVA2) was constructed as described previously [6]. The
GST fusion proteins were expressed in Escherichia coli and
purified by affinity chromatography on a GSH-agarose-4B
column, and this was followed by thrombin digestion to
obtain pure peptide fragments of VVA1-NTD-aH-C-D,
VVA1-CTD-aH-D¢-E¢ or VVA2 (HBS fragment). For the
competition assay, the various amounts of peptide compet-
itor were added to the mixture of VVA2 and VVA1 (at a
molar ratio of 2), and the mixture was incubated with
VVA1 beads at 37 °C for 30 min, and then washed and
eluted as described above. The eluent was separated by
SDS ⁄ PAGE and transferred to the polyvinylidene difluo-
ride membrane, and western blots were prepared using
anti-VVA2 as a standard protocol [6].
Acknowledgements
We would like to thank Professor Ta-Hsiu Liao for
his valuable suggestions, and Professor Zee-Fen
Chang, Laura Heraty and Dr Brett Hosking for their
critical reading of this manuscript. This work was sup-
ported in part by Grant NSC 89-2320-B-002-238 and
Grant NSC 93-2320-B-002-107 from the National
Science Council, Republic of China.

References
1 Lin JY, Jeng TW, Chen CC, Shi GY & Tung TC
(1973) Isolation of a new cardiotoxic protein from the
edible mushroom, Volvariella volvacea. Nature 246,
524–525.
2 Fassold E, Slade AM, Lin JY & Nayler WG (1976) An
effect of the cardiotoxic protein volvatoxin A on the
function and structure of heart muscle cells. J Mol Cell
Cardiol 8, 501–519.
3 Lin JY & Shi GY (1976) Toxicity of Volvatoxin A
isolated from edible mushroom, Volvariella volvacea.
J Taiwan Pharmaceut Assoc 28, 96–103.
4 Lin JY, Lin YJ, Chen CC, Wu HL, Shi GY & Jeng TW
(1974) Cardiotoxic protein from edible mushrooms.
Nature 252, 235–237.
5 Wu PT, Lin SC, Hsu CI & Lin JY (2004) Studies on
structure and function of VVA1 isolated from Volvar-
iella volvacea by site-directed mutagenesis. In Twelfth
Symposium on Recent Advances in Cellular and Molecu-
lar Biology (Chung-Ming, Chang, eds), February 2–4,
pp. 124. The Chinese Society of Cell and Molecular
Biology, Taipei, Taiwan.
6 Weng YP, Lin YP, Hsu CI & Lin JY (2004) Functional
domains of a pore-forming cardiotoxic protein, volva-
toxin A2. J Biol Chem 279, 6805–6814.
7 Walker B, Krishnasastry M, Zorn L & Bayley H (1992)
Assembly of the oligomeric membrane pore formed by
Staphylococcal alpha-hemolysin examined by truncation
mutagenesis. J Biol Chem 267, 21782–21786.
P T. Wu et al. VVA1 is a novel toxin regulator of VVA2

FEBS Journal 273 (2006) 3160–3171 ª 2006 The Authors Journal compilation ª 2006 FEBS 3169
8 Hotze EM, Heuck AP, Czajkowsky DM, Shao Z,
Johnson AE & Tweten RK (2002) Monomer–monomer
interactions drive the prepore to pore conversion of a
beta-barrel-forming cholesterol-dependent cytolysin.
J Biol Chem 277, 11597–11605.
9 Czajkowsky DM, Hotze EM, Shao Z & Tweten RK
(2004) Vertical collapse of a cytolysin prepore moves its
transmembrane beta-hairpins to the membrane. EMBO
J 23, 3206–3215.
10 Bayley H, Jayasinghe L & Wallace M (2005) Prepore
for a breakthrough. Nat Struct Mol Biol 12, 385–386.
11 Valeva A, Palmer M & Bhakdi S (1997) Staphylococcal
alpha-toxin: formation of the heptameric pore is
partially cooperative and proceeds through multiple
intermediate stages. Biochemistry 36, 13298–13304.
12 Song L, Hobaugh MR, Shustak C, Cheley S, Bayley H
& Gouaux JE (1996) Structure of staphylococcal alpha-
hemolysin, a heptameric transmembrane pore. Science
274, 1859–1866.
13 Choe S, Bennett MJ, Fujii G, Curmi PM, Kantardjieff
KA, Collier RJ & Eisenberg D (1992) The crystal
structure of diphtheria toxin. Nature 357, 216–222.
14 Malovrh P, Viero G, Serra MD, Podlesek Z, Lakey JH,
Macek P, Menestrina G & Anderluh G (2003) A novel
mechanism of pore formation: membrane penetration
by the N-terminal amphipathic region of equinatoxin.
J Biol Chem 278, 22678–22685.
15 Milne JC, Furlong D, Hanna PC, Wall JS & Collier RJ
(1994) Anthrax protective antigen forms oligomers

during intoxication of mammalian cells. J Biol Chem
269, 20607–20612.
16 Benson EL, Huynh PD, Finkelstein A & Collier RJ
(1998) Identification of residues lining the anthrax
protective antigen channel. Biochemistry 37, 3941–3948.
17 Barth H, Aktories K, Popoff MR & Stiles BG (2004)
Binary bacterial toxins: biochemistry, biology, and
applications of common Clostridium and Bacillus
proteins. Microbiol Mol Biol Rev 68, 373–402.
18 Montecucco C (2001) Detoxification of a bacterial
toxin by the toxin itself. Trends Pharmacol Sci 22,
493–494.
19 Liu CL, Tsai CC, Lin SC, Wang LI, Hsu CI, Hwang
MJ & Lin JY (2000) Primary structure and function
analysis of the Abrus precatorius agglutinin A chain by
site-directed mutagenesis. Pro(199) of amphiphilic
alpha-helix H impairs protein synthesis inhibitory activ-
ity. J Biol Chem 275, 1897–1901.
20 Patel HV, Vyas AA, Vyas KA, Liu YS, Chiang CM,
Chi LM & Wu W (1997) Heparin and heparan sulfate
bind to snake cardiotoxin. Sulfated oligosaccharides as
a potential target for cardiotoxin action. J Biol Chem
272, 1484–1492.
21 Vyas AA, Pan JJ, Patel HV, Vyas KA, Chiang CM,
Sheu YC, Hwang JK & Wu W (1997) Analysis of bind-
ing of cobra cardiotoxins to heparin reveals a new
beta-sheet heparin-binding structural motif. J Biol Chem
272, 9661–9670.
22 Gatineau E, Takechi M, Bouet F, Mansuelle P, Rochat
H, Harvey AL, Montenay-Garestier T & Menez A

(1990) Delineation of the functional site of a snake
venom cardiotoxin: preparation, structure, and function
of monoacetylated derivatives. Biochemistry 29, 6480–
6489.
23 Venema RC, Sayegh HS, Arnal JF & Harrison DG
(1995) Role of the enzyme calmodulin-binding domain
in membrane association and phospholipid inhibition of
endothelial nitric oxide synthase. J Biol Chem 270,
14705–14711.
24 Lin SC, Lo YC, Lin JY & Liaw YC (2004) Crystal
structures and electron micrographs of fungal volva-
toxin A2. J Mol Biol 343, 477–491.
25 Eisenberg D, Weiss RM & Terwilliger TC (1982) The
helical hydrophobic moment: a measure of the amphi-
philicity of a helix. Nature 299, 371–374.
26 Hsu YT & Youle RJ (1997) Nonionic detergents induce
dimerization among members of the Bcl-2 family. J Biol
Chem 272, 13829–13834.
27 Bhakdi S, Fussle R & Tranum-Jensen J (1981) Staphy-
lococcal alpha-toxin: oligomerization of hydrophilic
monomers to form amphiphilic hexamers induced
through contact with deoxycholate detergent micelles.
Proc Natl Acad Sci USA 78, 5475–5479.
28 Ole JB, James HG, Thorkild CB-H & Jesper BH (1982)
Electroimmunochemical analysis of amphiphilic proteins
and glycolipids stained with Sudan Black-containing
detergent micelles. Electrophoresis 3, 89–98.
29 Tone B & Ole JB (1986) Electrophoretic migration velo-
city of amphiphilic proteins increases with decreasing
Triton X-100 concentration: a new characteristic for

their identification. Electrophoresis 7, 197–203.
30 Panchal RG, Cusack E, Cheley S & Bayley H (1996)
Tumor protease-activated, pore-forming toxins from a
combinatorial library. Nat Biotechnol 14, 852–856.
31 Liu S, Aaronson H, Mitola DJ, Leppla SH & Bugge
TH (2003) Potent antitumor activity of a urokinase-acti-
vated engineered anthrax toxin. Proc Natl Acad Sci
USA 100, 657–662.
32 Johannes L & Decaudin D (2005) Protein toxins: intra-
cellular trafficking for targeted therapy. Gene Ther 12,
1360–1368.
33 Panchal RG, Smart ML, Bowser DN, Williams DA &
Petrou S (2002) Pore-forming proteins and their applica-
tion in biotechnology. Curr Pharm Biotechnol 3, 99–115.
34 Michl P & Gress TM (2004) Bacteria and bacterial tox-
ins as therapeutic agents for solid tumors. Curr Cancer
Drug Targets 4, 689–702.
35 Potrich C, Tomazzolli R, Dalla Serra M, Anderluh G,
Malovrh P, Macek P, Menestrina G & Tejuca M (2005)
Cytotoxic activity of a tumor protease-activated pore-
forming toxin. Bioconjug Chem 16, 369–376.
VVA1 is a novel toxin regulator of VVA2 P T. Wu et al.
3170 FEBS Journal 273 (2006) 3160–3171 ª 2006 The Authors Journal compilation ª 2006 FEBS
36 Panchal RG (1998) Novel therapeutic strategies to selec-
tively kill cancer cells. Biochem Pharmacol 55, 247–252.
37 Liu S, Bugge TH & Leppla SH (2001) Targeting of
tumor cells by cell surface urokinase plasminogen acti-
vator-dependent anthrax toxin. J Biol Chem 276, 17976–
17984.
38 Liu S, Redeye V, Kuremsky JG, Kuhnen M, Molinolo

A, Bugge TH & Leppla SH (2005) Intermolecular com-
plementation achieves high-specificity tumor targeting
by anthrax toxin. Nat Biotechnol 23, 725–730.
39 Rono B, Romer J, Liu S, Bugge TH, Leppla SH &
Kristjansen PE (2006) Antitumor efficacy of a urokinase
activation-dependent anthrax toxin. Mol Cancer Ther 5,
89–96.
40 Kaku H, Tanaka Y, Tazaki K, Minami E, Mizuno H &
Shibuya N (1996) Sialylated oligosaccharide-specific
plant lectin from Japanese elderberry (Sambucus sie-
boldiana) bark tissue has a homologous structure to type
II ribosome-inactivating proteins, ricin and abrin.
cDNA cloning and molecular modeling study. J Biol
Chem 271, 1480–1485.
41 Wu YH, Shih SF & Lin JY (2004) Ricin triggers apop-
totic morphological changes through caspase-3 cleavage
of BAT3. J Biol Chem 279, 19264–19275.
42 Shih SF, Wu YH, Hung CH, Yang HY & Lin JY (2001)
Abrin triggers cell death by inactivating a thiol-specific
antioxidant protein. J Biol Chem 276, 21870–21877.
43 Wilchek M, Miron T & Kohn J (1984) Affinity chroma-
tography. Methods Enzymol 104, 3–55.
Supplementary material
The following supplementary material is available
online:
Table S1. Sequences of the primers used for cloning
volvatoxin A1 (VVA1) cDNA and peptide competi-
tors.
Fig. S1. Nucleotide and deduced protein sequence of
volvatoxin A1 (VVA1). The nucleotide sequence and

deduced amino acid sequence of VVA1 cDNA (Gen-
Bank accession number AY952461). The ORF consists
of 393 amino acid residues. The first nucleotide and
amino acid of VVA1 are boxed. Nucleotide and amino
acid residues are numbered on the left and right,
respectively. Arrows mark the primers used in RACE-
PCR. The asterisk denotes the stop codon at nucleo-
tide position 1180. The amino acid sequence deduced
from VVA1 cDNA is shown by a one-letter code.
Fig. S2. Schematic representation of a structural model
of volvatoxin A1 (VVA1). The figure shows a ribbon
plot representation of the proposed models for VVA1.
b-Sheets are in yellow, a-helices are in pink, and the
loops are in white–blue. Each VVA1 domain was
modeled by swissmodel ()
using the crystal structure of volvatoxin A2 (VVA2)
(PDBID:1PP0) (J Mol Biol 343, 477–491, 2004) as a
template. Two domains were linked manually by pro-
gram O.
Fig. S3. Effects of detergents and liposomes on the
oligomerization of volvatoxin A2 (VVA2). (A) Triton
X-100 induced the oligomerization of VVA2. VVA2
(45 nm) was incubated with various detergents in
50 mm Tris, pH 8.0, at 37 °C for 30 min. The forma-
tion of VVA2 oligomer was determined by 10%
SDS ⁄ PAGE analysis. (B) Oligomerization of VVA2
was induced by liposomes. After treatment of VVA2
with or without 5 mm liposomes in NaCl ⁄ P
i
, the reac-

tion products were analyzed by 10% SDS ⁄ PAGE.
Fig. S4. Anti-volvatoxin A2 (VVA2) IgG. VVA2 and
volvatoxin A1 (VVA1) were analyzed by 10%
SDS ⁄ PAGE, followed by western blots and anti-VVA2
IgG.
This material is available as part of the online article
from
P T. Wu et al. VVA1 is a novel toxin regulator of VVA2
FEBS Journal 273 (2006) 3160–3171 ª 2006 The Authors Journal compilation ª 2006 FEBS 3171