Báo cáo khoa học: The A-subunit of surface-bound Shiga toxin stimulates clathrin-dependent uptake of the toxin ppt

The A-subunit of surface-bound Shiga toxin stimulates
clathrin-dependent uptake of the toxin
Maria L. Torgersen
1
, Silje U. Lauvrak
1
and Kirsten Sandvig
1,2
1 Institute for Cancer Research, The Norwegian Radium Hospital, Montebello, Oslo, Norway
2 Department of Molecular Biosciences, University of Oslo, Norway
The Shiga toxins consist of one enzymatically active
A-subunit noncovalently linked to a stable pentamer
of binding subunits (StxB), which in most cases bind
specifically to the glycosphingolipid Gb3 in the plasma
membrane [1]. To exert its toxic effect, the holotoxin
must be endocytosed and retrogradely transported via
the Golgi apparatus to the endoplasmic reticulum,
where the A-subunit is translocated to the cytosol and
inhibits protein synthesis [2,3]. The Shiga family of
toxins is divided into two main groups, based on
antigenic differences. Shiga toxin (Stx) produced by
Shigella dysenteriae and Shiga toxin 1 (Stx1) secreted
by certain strains of Eshcerichia coli (Shiga toxin pro-
ducing E. coli; STEC) are virtually identical and differ
only in one amino acid. The toxins in the Shiga toxin
2 family (Stx2, Stx2c, Stx2d and Stx2e) are also secre-
ted by STEC and have a similar structure to the toxins
in the first group, but differ both functionally and
immunologically [1,4,5]. To understand the differences
between the effects of the various Stx, it is important
to clarify the uptake mechanisms for these toxins.

It has been shown that Stx is internalized via cla-
thrin-dependent endocytosis in several cell types,
although there is evidence that also clathrin-independ-
ent mechanisms are partly involved [2,6–8]. Import-
antly, several studies have been concerned with the
uptake and intracellular transport of the Stx B-subunit
[8–21] but so far, investigations of a possible role of
Keywords
clathrin; endocytosis; Shiga toxin
Correspondence
K. Sandvig, Institute for Cancer Research,
The Norwegian Radium Hospital,
Montebello, 0310 Oslo, Norway
Fax: +47 22508692
Tel: +47 22934294
E-mail:
(Received 12 May 2005, revised 14 June
2005, accepted 24 June 2005)
doi:10.1111/j.1742-4658.2005.04835.x
Shiga toxin can be internalized by clathrin-dependent endocytosis in differ-
ent cell lines, although it binds specifically to the glycosphingolipid Gb3. It
has been demonstrated previously that the toxin can induce recruitment of
the toxin–receptor complex to clathrin-coated pits, but whether this process
is concentration-dependent or which part of the toxin molecule is involved
in this process, have so far been unresolved issues. In this article, we show
that the rate of Shiga toxin uptake is dependent on the toxin concentration
in several cell lines [HEp-2, HeLa, Vero and baby hamster kidney (BHK)],
and that the increased rate observed at higher concentrations is strictly
dependent on the presence of the A-subunit of cell surface-bound toxin.
Surface-bound B-subunit has no stimulatory effect. Furthermore, this

increase in toxin endocytosis is dependent on functional clathrin, as it did
not occur in BHK cells after induction of antisense to clathrin heavy chain,
thereby blocking clathrin-dependent endocytosis. By immunofluorescence,
we show that there is an increased colocalization between Alexa-labeled
Shiga toxin and Cy5-labeled transferrin in HeLa cells upon addition of
unlabeled toxin. In conclusion, the data indicate that the Shiga toxin
A-subunit of cell surface-bound toxin stimulates clathrin-dependent uptake
of the toxin. Possible explanations for this phenomenon are discussed.
Abbreviations
BHK, baby hamster kidney; CHC, clathrin heavy chain; CT, cholera toxin; MESNa, mercaptoethanesulfonic acid; STEC, Shiga toxin producing
Escherichia coli; Stx, Shiga toxin; Stx1, Shiga toxin 1; Stx2, Shiga toxin 2; StxB, Shiga toxin B subunit.
FEBS Journal 272 (2005) 4103–4113 ª 2005 FEBS 4103
the A-subunit for uptake and intracellular routing of
the toxin have been few. By electron microscopy stud-
ies, intact Stx has been shown to preferentially localize
to clathrin-coated pits in HeLa cells [22,23], and both
acidification and potassium depletion of cytosol, which
inhibit clathrin-dependent endocytosis, have been
shown to protect the cells against Stx [22–24]. Further-
more, induced expression of antisense to clathrin heavy
chain (CHC) in a baby hamster kidney (BHK) cell line
[25–27], thereby blocking the clathrin-dependent endo-
cytosis, also protects against Stx [2] and reduces the
endocytic uptake of the toxin [8]. By using this BHK
cell line, we could quantify the initial Stx internalizat-
ion, both when clathrin-dependent endocytosis was
operating and when it was blocked by antisense
expression. Toxin uptake was decreased by about 50%
upon inhibition of clathrin-dependent endocytosis, but
it should be noted that a block in clathrin-dependent

endocytosis might lead to increased toxin uptake via
other mechanisms.
Importantly, Stx seems to be able to induce its own
entry from clathrin-coated pits. Electron microscopy
studies revealed that bound Stx is evenly distributed
on the cell surface at low temperature, while after
shifting to 37 °C, the toxin is aggregated in clathrin-
coated pits [22,23]. There is no obvious explanation as
to how the Stx–Gb3 complex is recruited to clathrin-
coated pits, given that the receptor is a glycosphingo-
lipid and not a protein with specific sequences required
for interaction with the sorting machinery. One pos-
sible explanation is that the Stx–Gb3 complex interacts
with another protein that is internalized via clathrin-
dependent endocytosis. In fact, it has been shown by
crosslinking experiments that members of the Shiga
family of toxins can interact with 27 and 40 kDa mole-
cules at the cell surface of Vero and CaCo2 cells [28],
suggesting that interaction with accessory proteins
might facilitate clathrin-mediated uptake. However,
other possible explanations do exist. The toxin could
induce toxin-specific signaling leading to recruitment
of clathrin. Furthermore, crosslinking of Gb3 and per-
haps of lipid rafts might be important for this process.
Interesting in this connection, is the finding that the
epidermal growth factor receptor present in lipid rafts
recruits clathrin to the membrane upon binding of epi-
dermal growth factor [29].
We here demonstrate that the rate of Stx uptake is
up-regulated by increasing toxin concentrations by a

mechanism in which the A-subunit of the surface-
bound toxin is required. Importantly, this increased
endocytosis is clathrin-dependent, and it seems to be
caused by an increased recruitment of the toxin–recep-
tor complex to clathrin-coated pits.
Results
Stx stimulates its own rapid entry in several
cell lines
We have shown previously that Stx, by a so far
unknown mechanism, is able to mediate redistribution
of the Stx–Gb3 complex to clathrin-coated pits, before
being internalized [22]. In order to investigate whether
Stx internalization is dependent on the amount of toxin
that is bound to the membrane, different cell types were
incubated with increasing concentrations of unlabeled
Stx or Stx1 (0–20 nm) in addition to a constant concen-
tration of TAG- and biotin-labeled Stx (0.33 nm), before
the amount of internalized Stx was measured and com-
pared to untreated control (no unlabeled Stx added). As
shown in Fig. 1A, treating cells with increasing concen-
trations of unlabeled Stx1 increased the fraction of inter-
nalized TAG-labeled Stx in nearly all the cell lines tested
(BHK, Vero, HEp-2 and HeLa) by a factor of  2 dur-
ing a 10 min incubation at 37 °C. Thus, Stx seemed to
have a stimulatory effect on its own uptake in these cells.
In one cell line tested, A431, only a slight stimulatory
effect was observed.
To investigate whether this stimulatory effect was
affected by unbound toxin in the medium, increasing
concentrations of Stx1 (0–20 nm) and a constant con-

centration of TAG- and biotin-labeled Stx (0.33 nm)
were prebound to HEp-2 cells for 30 min on ice. Then
the medium was collected and warmed to 37 °C, while
the cells were washed twice. Endocytic uptake at 37 °C
was measured for 10 min in the presence or absence of
the warmed preincubation medium (i.e., with or with-
out unbound toxin in the solution), and the results are
compared in Fig. 1B. Clearly, the presence of unbound
toxin in the medium does not affect the stimulation of
toxin uptake.
To study the actual rate of Stx uptake, and more spe-
cifically to compare the rate of control and stimulated
Stx uptake, TAG- and biotin-labeled Stx (0.33 nm) was
added to HEp-2 cells, and internalized toxin was meas-
ured after different time periods in the presence or
absence of a stimulatory dose of unlabeled Stx1 (13 nm).
As shown in Fig. 2A, the internalization of Stx is quite
rapid both at high and low toxin concentrations, but less
toxin is internalized at low toxin concentrations even
during the first 10 min of incubation. The fact that the
toxin accumulation leveled off at later time points may
be due to recycling. For easy comparison, the endocytic
control values obtained in the absence of unlabeled Stx1
were normalized to 100% (Fig. 2B). The data show that
whereas no significant difference was seen after 2.5 min,
there was a clear difference after 5 min and no further
Endocytosis of Shiga toxin M. L. Torgersen et al.
4104 FEBS Journal 272 (2005) 4103–4113 ª 2005 FEBS
increase after 7.5 min. Thus, the stimulation of Stx
uptake is a rapid process.

The stimulation of Stx uptake is specific for Stx
and requires surface-bound toxin A-subunit
In order to investigate whether the stimulation of
toxin uptake at high toxin concentrations was an
effect specific for Stx, experiments using cholera toxin
(CT) were also conducted. A possible cross-stimulation
between CT and Stx was investigated because both
toxins bind to glycosphingolipids in the plasma mem-
brane (Gb3 and GM1, respectively), and might be
localized in the same microdomain. Increasing the
toxin concentrations would presumably give increased
crosslinking and aggregation of toxin–receptor com-
plexes, and the possibility existed that the two toxins
would be endocytosed together in these aggregates.
Importantly, Stx uptake was unchanged upon addition
of increasing concentrations of CT (0–20 nm)as
shown in Fig. 3A. Also, endocytosis of TAG- and bio-
tin-labeled CT (0.12 nm) was measured in HEp-2 cells
incubated with increasing concentrations of unlabeled
Stx1 (0–20 nm). There was no increase in CT uptake
under these conditions (data not shown). Furthermore,
endocytosis of TAG- and biotin-labeled CT (0.12 nm)
was measured at increasing CT concentrations (0–
20 nm) in A431 cells. There was no stimulation of CT
endocytosis with increasing toxin concentrations (data
not shown), which is in accordance with data obtained
previously for CT uptake in BHK cells [30]. Thus, sti-
mulation of toxin uptake at high toxin concentrations
seems to be an effect specific for Stx. We next investi-
gated whether Stx2, which also binds to Gb3, could

stimulate Stx uptake. Stx1 and Stx2 show 55% and
57% amino acid similarity in the A- and B-chain,
respectively [31], but despite this similarity they are
immunologically distinct. Endocytosis of Stx was
measured upon addition of increasing amounts of
unlabeled Stx2 (0–20 nm), and as shown in Fig. 3A,
only at very high concentrations of Stx2 was a slight
stimulation of Stx uptake observed.
To explore the mechanism of this specific stimulation
of Stx uptake, we wanted to investigate whether an
aggregation of Gb3 in the membrane, mediated by the
pentameric B-subunit, would be sufficient to increase
the toxin uptake. To this end, endocytosis of TAG- and
biotin-labeled Stx (0.33 nm) was compared in cells
Unlabeled Stx1 added (nM)
0 5 10 15 20
Shiga toxin endocytosis
(compared to untreated control)
0.5
1.0
1.5
2.0
2.5
HEp-2
Vero
BHK
HeLa
A431
Unlabeled Stx1 added (nM)
0 5 10 15 20

Shiga toxin endocytosis
(compared to untreated control)
0.0
0.5
1.0
1.5
2.0
2.5
3.0
Stx1 removed
Stx1 present
A
B
Fig. 1. Endocytosis of Stx at increasing toxin concentrations. (A)
Stimulation of Stx uptake in different cell lines. The cells were pre-
incubated with increasing concentrations of unlabeled Stx1
(0–20 n
M) in addition to a constant concentration of TAG- and bio-
tin-labeled Stx (0.33 n
M) in Hepes buffered MEM at 0 °C for
30 min. Then the cells were washed, and the incubation continued
for 10 min at 37 °C, except for BHK cells, which were incubated
for 20 min. Bound and endocytosed toxin were quantified as des-
cribed in Experimental procedures. The endocytic values obtained
upon Stx stimulation were compared to control values (no unlabe-
led Stx1 added) in one representative experiment from each cell
line (mean values ± deviations between duplicates). The experi-
ment was repeated separately 3–6 times for every cell line. (B)
Uptake of Stx with or without unbound Stx1 in solution. HEp-2 cells
were preincubated with TAG- and biotin-labeled Stx (0.33 n

M)in
addition to increasing concentrations of Stx1 (0–20 n
M) for 30 min
at 0 °C. Then the medium was collected and warmed, and the cells
were washed, before endocytosis was measured at 37 °C for
10 min either in the presence of the warmed preincubation med-
ium (i.e., with unbound toxin present) or with warmed toxin-free
medium. The endocytic values are compared to untreated control
(no unlabeled Stx1 added at 0 °C) (mean values ± deviations
between two independent experiments, each done in duplicate).
M. L. Torgersen et al. Endocytosis of Shiga toxin
FEBS Journal 272 (2005) 4103–4113 ª 2005 FEBS 4105
treated with either increasing concentrations of Shiga
holotoxin or Stx B-subunit (0–20 nm) in parallel experi-
ments, in both HEp-2 and HeLa cells. As shown in
Fig. 3B, there was no stimulatory effect on Stx endo-
cytosis upon addition of Stx B-subunit alone in either
cell type; thus, the stimulation seems to require the pres-
ence of the Stx A-subunit.
The self-stimulated Stx uptake is
clathrin-dependent
In the internalization of Stx, both clathrin-dependent
and clathrin-independent mechanisms seem to be
Time (min)
0 10203060
Shiga toxin endocytosis
(% of total cell-associated toxin)
0
10
20

30
40
50
A
B
Time (min)
Shiga toxin endocytosis
(% of untreated control)
0
50
100
150
200
250
2.5 5 7.5 10 30 60
Fig. 2. Uptake of Stx at high or low toxin concentration with time.
(A) HEp-2 cells were incubated with (s) or without (d) unlabeled
Stx1 (13 n
M) in addition to TAG- and biotin-labeled Stx (0.33 nM)in
Hepes buffered MEM at 37 °C for 2.5–60 min, before bound and
endocytosed Stx were quantified as described in Experimental pro-
cedures (mean values ± deviations between two independent
experiments, each done in duplicate). (B) Control values (filled bars)
were normalized to 100% for comparison to the endocytic values
obtained upon Stx1 stimulation (open bars).
Unlabeled Stx or StxB added (nM)
0 5 10 15 20
Shiga toxin endocytosis
(% of total cell-associated toxin)
0

10
20
30
40
50
60
HEp-2 Stx
HEp-2 StxB
HeLa Stx
HeLa StxB
Unlabeled toxin added (nM)
0 5 10 15 20
Shiga toxin endocytosis
(compared to untreated control)
0.0
0.5
1.0
1.5
2.0
2.
A
B
5
Stx1
Stx2
CT
Fig. 3. Endocytosis of Stx in cells incubated with increasing con-
centrations of Stx, Stx1, Stx2, CT or Stx B-subunit. (A) Effect of
surface-bound Stx1, Stx2 or CT. HEp-2 cells were preincubated
with TAG- and biotin-labeled Stx (0.33 n

M) in addition to increasing
concentrations of Stx1, Stx2 or CT (0–20 n
M) in Hepes buffered
MEM for 30 min at 0 °C. Then the cells were washed, and the
incubation continued for 10 min at 37 °C, before bound and endo-
cytosed toxin were quantified as described in Experimental proce-
dures. The endocytic values obtained were compared to untreated
control values (no unlabeled toxin added) in a representative
experiment (mean values ± deviations between duplicates). Each
experiment was repeated separately 3–4 times. (B) Effect of sur-
face-bound Stx or StxB. Increasing concentrations of Shiga holo-
toxin or Stx B-subunit, both at 0–20 n
M, were added to HEp-2 or
HeLa cells in Hepes buffered MEM in addition to a constant con-
centration of TAG- and biotin-labeled Stx (0.33 n
M). The cells were
incubated for 10 min at 37 °C, before bound and endocytosed toxin
were quantified as described in Experimental procedures. The
amount of internalized toxin is presented as percent of total cell-
associated toxin (mean values ± deviations between duplicates in a
representative experiment.) Stimulation with the holotoxin or the
B-subunit was assayed in parallel in four independent experiments
in each cell type.
Endocytosis of Shiga toxin M. L. Torgersen et al.
4106 FEBS Journal 272 (2005) 4103–4113 ª 2005 FEBS
involved in different cell types [2,6–8]. To further
elucidate the role of clathrin in Stx uptake, and
more specifically, to study the involvement of clath-
rin in the self-stimulated uptake of Stx, a BHK cell
line with inducible expression of antisense to CHC

was used. Upon induction, the clathrin-dependent
endocytosis is completely inhibited in these cells.
After two days of expression of antisense to CHC,
uptake of TAG- and biotin-labeled Stx (0.33 nm)
was measured, and as shown in Fig. 4, when clathrin-
dependent endocytosis was inhibited, Stx endocyto-
sis was reduced by 50% when no unlabeled toxin
was added; confirming previously published data on
clathrin-dependent uptake of Stx in these cells [8].
Interestingly, as shown in Fig. 4, the presence of
increasing concentrations of unlabeled Stx1 (0–20 nm)
lead to a marked increase in Stx uptake only in cells
where the clathrin-dependent uptake was functional. In
contrast, toxin endocytosis was not increased at high
toxin concentrations after induction of antisense CHC.
This suggests that high toxin concentrations preferen-
tially stimulate the clathrin-dependent fraction of Stx
uptake. To rule out the possibility that high Stx con-
centrations lead to a stimulation of clathrin-dependent
uptake in general, the endocytosis of transferrin was
measured in the presence of increasing Stx1 concentra-
tions (0–20 nm). Under these conditions, the transfer-
rin uptake remained unchanged (data not shown). In
addition, uptake of the plant toxin ricin, which binds
to both glycoproteins and glycolipids with terminal
galactose, and which presumably is internalized by dif-
ferent endocytic mechanisms, was not affected by high
concentrations of Stx1 (data not shown).
Alexa-Stx is redistributed to transferrin-containing
domains upon addition of unlabeled Stx1

Increasing amounts of Stx1 might lead to aggregation
of Stx–Gb3 complexes into large clusters that subse-
quently could increase the uptake efficiency of the
toxin in a manner where apparently also the A-subunit
plays a role. Because stimulating the cells with increas-
ing concentrations of Stx1 does not increase the clath-
rin-dependent endocytosis in general, the increase in
Stx uptake that we observe might be explained by an
increased recruitment of Stx–Gb3 complexes to clath-
rin-coated pits. To explore this, HeLa cells grown on
coverslips were treated with Alexa-labeled Stx (1.3 nm)
and Cy5-labeled transferrin (130 nm) as a marker of
clathrin-coated pits, and incubated with or without
unlabeled Stx1 (20 nm). Then single plane sections
were made by confocal microscopy and the fluores-
cence intensities were quantified. Earlier results from
electron microscopy indicate that Stx is evenly distri-
buted at the cell surface of HeLa cells at 0 °C, and
there is a redistribution of the toxin to clathrin-coated
pits only at higher temperatures [22]. In agreement
with these data there was a diffuse labeling of HeLa
cells incubated with Alexa-labeled Stx at 0 °C (data
not shown). To allow for slow redistribution of toxin,
the temperature was raised to 10 °C, and colocaliza-
tion of Alexa-Stx and Cy5-transferrin was quantified
and compared in cells treated with or without unlabe-
led Stx1. As shown in Fig. 5, incubation of the cells
with unlabeled Stx1 (in addition to Alexa-Stx) lead to
a marked increase in colocalization of Alexa-Stx and
Cy5-transferrin, indicating an increased redistribution

of Stx to transferrin-containing microdomains ⁄ clath-
rin-coated pits at the plasma membrane.
Recent data indicate a correlation between the level
of the proteoglycan heparan sulfate on the cell surface
and the rate of prion protein endocytosis [32]. In order
to investigate whether proteoglycans are involved in
Stx endocytosis, HEp-2 cells were treated with chlor-
ate, which inhibits the sulfation of proteoglycans [33],
or treated with heparinase I, which cleaves off surface-
bound heparan sulfate [34]. Neither treatment reduced
the endocytic uptake of Stx (data not shown). Thus,
Unlabeled Stx1 added (nM)
Shiga toxin endocytosis
(% of total cell-associated toxin)
0
10
20
30
40
50
60
Non-induced cells
Induced cells
01020
Fig. 4. Endocytosis of Stx in noninduced and induced BHK cells
with inducible expression of antisense CHC, which inhibits clathrin-
dependent endocytosis. After two days of induction, the cells were
washed and preincubated with increasing concentrations of unlabe-
led Stx1 (0–20 n
M) in addition to TAG- and biotin-labeled Stx

(0.33 n
M) in Hepes buffered MEM at 0 °C for 30 min. Then the
cells were washed, and the incubation continued for 20 min at
37 °C, before bound and endocytosed toxin were quantified as des-
cribed in Experimental procedures. The amount of internalized toxin
is presented as percent of total cell-associated toxin in noninduced
cells (filled bars) and induced cells (open bars) (mean values ± devi-
ations between duplicates) in a representative experiment. The
experiment was repeated separately four times.
M. L. Torgersen et al. Endocytosis of Shiga toxin
FEBS Journal 272 (2005) 4103–4113 ª 2005 FEBS 4107
there is no evidence for involvement of proteoglycans
in Stx internalization. Furthermore, to investigate whe-
ther the charge of Stx is of importance for the toxin
uptake, internalization of Stx was measured at differ-
ent pH values, ranging from 5.5 to 8.5, after different
time points. However, under these conditions the rate
of Stx uptake was unchanged (data not shown).
Discussion
In the present article, we show that Stx is able to sti-
mulate its own entry by a process dependent on the
toxin concentration, the surface-binding of the com-
plete AB
5
toxin and clathrin-dependent endocytosis.
This self-stimulatory effect of Stx uptake seems to be
mediated by an increased recruitment of the Stx–Gb3
complex to clathrin-coated pits.
Stx can be internalized by clathrin-dependent endo-
cytosis, despite the lipid nature of its receptor Gb3

[2,4,7,22,35]. To study this process we decided to
investigate whether Stx uptake was dependent on the
toxin concentration added to the cells, and indeed this
seemed to be the case. When increasing concentrations
of unlabeled Stx were added to different cell lines
(HEp-2, BHK, Vero, HeLa and A431 cells), uptake of
labeled Stx was increased approximately twofold in all
the cell lines except A431 cells. This self-stimulated Stx
uptake was found to be mediated by clathrin-dependent
endocytosis, as suggested from experiments using BHK
cells with inducible expression of antisense to CHC.
Based on this result, one might speculate whether the
difference in stimulation of Stx uptake between cell
lines is due to differential involvement of clathrin-
dependent endocytosis of Stx in the cells. Interestingly,
No unlabeled Stx1 High unlabeled Stx1
Fig. 5. Localization of Alexa-labeled Stx and
Cy5-labeled transferrin in cells treated with
or without unlabeled Stx1. HeLa cells were
incubated with Alexa-labeled Stx (1.3 n
M)
and Cy-5 labeled transferrin (130 n
M), and
one half of the cells was in addition incuba-
ted with 20 n
M of Stx1 (high unlabeled Stx1).
All the cells were incubated for 40 min at
10 °C before fixation. Quantification of the
extent of colocalization between Stx and
transferrin (as percentage of total amount of

Stx) when the cells were incubated without
unlabeled Stx1 (black bars, n ¼ 29) or with
high unlabeled Stx1 (grey bars, n ¼ 43) is
shown in the lower panel. Data were pooled
from four independent experiments.
Endocytosis of Shiga toxin M. L. Torgersen et al.
4108 FEBS Journal 272 (2005) 4103–4113 ª 2005 FEBS
there were large differences in the actual amount of Stx
internalized in each cell line after 10 min at low toxin
concentrations, ranging from 13% in A431 cells to
33% in BHK cells (internalized toxin as percent of total
cell-associated toxin; data not shown). In general,
clathrin-dependent endocytosis is much faster than
clathrin-independent mechanisms studied so far, due to
the  1 min half-life of clathrin-coated pits at the cell
surface [36]. Thus, a high rate of toxin uptake might
reflect that a larger fraction of toxin endocytosis is
clathrin-dependent, even at low toxin concentrations.
Upon stimulation of Stx endocytosis with increasing
concentrations of toxin, clathrin-dependent and -inde-
pendent endocytosis were not increased in general.
There was no increase in the uptake of transferrin or
ricin, rather the Stx uptake seemed to be increased spe-
cifically. This could be explained by an increased local-
ization of Stx–Gb3 complexes in clathrin-coated pits,
and was investigated by immunofluorescence using
Alexa-labeled Stx (in the absence and presence of
unlabeled toxin) and Cy5-labeled transferrin as a mar-
ker of clathrin-coated pits. Although Alexa-Stx gave a
heterogeneous labeling of the cells, as reported pre-

viously by others [9], the results clearly showed an
increased colocalization of Stx and transferrin in cells
incubated with unlabeled Stx1 compared to control
cells incubated with Alexa-Stx only. The increased co-
localization of Stx and transferrin did not result from
an increased amount of transferrin receptor on the
plasma membrane, because neither the binding nor the
endocytosis of transferrin was increased upon incuba-
tion with high concentrations of Stx1 (data not
shown). Thus, it seems likely that Stx–Gb3 localization
in transferrin-containing clathrin-coated pits is
increased upon stimulation with high concentrations of
Stx1. These new results can explain our previously
published electron microscopy data regarding Stx
redistribution to clathrin-coated pits upon shifting the
temperature from 0 °Cto37°C [22]. In those experi-
ments, Stx at a concentration as high as 133 nm was
added to HeLa cells, and in agreement with the data
in the present article, a high degree of Stx redistribu-
tion to clathrin-coated pits was observed by electron
microscopy upon shifting the temperature.
It seems that only cell surface-bound toxin molecules
are responsible for the self-stimulation of Stx uptake,
as was shown by comparing the stimulation of toxin
uptake in the presence or absence of free toxin mole-
cules in the medium. This could be due to the much
higher local toxin concentration at the cell surface
compared to that in the medium. Calculations show
that there is  10 000 times higher concentration of
labeled Stx at the cell periphery than in the rest of the

medium. (For the assumptions involved, see Experi-
mental procedures.)
Importantly, binding of the complete AB
5
Stx struc-
ture seems to be specifically required for the ability of
high toxin concentrations to induce an increased rate
of toxin endocytosis. Control experiments using CT
showed that the observed stimulation of Stx endocyto-
sis was not due to an unspecific aggregation of gly-
cosphingolipids ⁄ lipid rafts in the plasma membrane,
nor was a mere aggregation of Gb3 by Stx B-subunits
sufficient to stimulate Stx internalization. Thus, the
presence of the A-subunit of the Shiga holotoxin seems
to be crucial for the self-stimulated uptake to occur. In
order to explain this, one might envision that the
A-subunits are directly involved in important interac-
tions either to other A-subunits, which could possibly
cluster the toxins, or to other plasma membrane pro-
teins, which might facilitate toxin internalization.
Alternatively, the A-subunit might influence the toxin
internalization indirectly by affecting the exposure of
its associated B-subunits. This could change the sur-
face location of the toxin or facilitate interactions with
other membrane proteins that might induce toxin
internalization. Also, as mentioned above, both toxin
subunits (A and ⁄ or B) might induce signaling that
could mediate toxin internalization, and this toxin-
induced signaling might differ, depending on whether
the intact toxin or the B-subunit bind to Gb3. How-

ever, these signaling pathways are largely unknown.
Interestingly, increasing concentrations of Stx2, which
also binds to Gb3 and has an A-subunit with 55%
amino acid similarity to Stx1, gave only a slight stimula-
tion of Stx uptake in HEp-2 cells, a further demonstra-
tion of the specificity of this process. Important in this
connection is that although Stx1 and Stx2 are structur-
ally very similar, the A-subunit of Stx2 has a different
orientation with respect to the B-subunits than in Stx1
[37]. This might influence the ability of Stx2 to stimulate
Stx1 uptake. The finding that Stx2 behaves differently
from Stx1 is in agreement with their different effects on
cells and in disease (for review see [4,5]).
The data shown in this article clearly reveal that
the Stx A-subunit bound to the cells via the B-pen-
tamer is responsible for the stimulation of endocyto-
sis. However, the requirement for the A-subunit
could be mediated via an effect of the B-moiety.
This illustrates the importance of detailed studies of
the role of the different subunits ⁄ toxin domains for
toxin uptake and intracellular transport. In order to
understand the action of Stx and the Shiga-like tox-
ins on the cellular level as well as in disease, each
step along the retrograde pathway from the cell sur-
face to the cytosol should be characterized.
M. L. Torgersen et al. Endocytosis of Shiga toxin
FEBS Journal 272 (2005) 4103–4113 ª 2005 FEBS 4109
Experimental procedures
Materials
Hepes, BSA, mercaptoethanesulphonic acid (MESNa),

n-octyl b-d-glucopyranoside, tetracycline and transferrin
were purchased from Sigma Chemical Co., St. Louis, MO,
USA. CT was obtained from Calbiochem, San Diego, CA,
USA. Stx was provided by JV Kozlov (Academy of Sciences
of Russia, Moscow, Russia), and by JE Brown (USAMRIID,
Fort Detrick, MD, USA), who also provided us with Shiga
toxin B-subunit. The plasmid encoding StxB-Sulf
2
was a kind
gift from B Goud (The Curie Institute, Paris, France). Stx1
and Stx2 were purchased from Toxin Technology, Sarasota,
FL, USA. Stx was labeled with Alexa from Molecular Probes
(Leiden, the Netherlands) according to the procedure given
by the company, and Cy5-labeled transferrin was purchased
from the same company.
Cells
HEp-2, HeLa, Vero and A431 cells were grown in DMEM
supplemented with 10% (v ⁄ v) fetal bovine serum, 100
unitsÆmL
)1
penicillin and 100 lgÆmL
)1
streptomycin. HEp-2
cells were seeded out in 24-well plates at a density of 5 · 10
4
cells per well one day prior to experiments, while HeLa and
Vero cells were seeded at 2 · 10
4
cells per well two days prior
to experiments. A431 cells were seeded at 5 · 10

4
cells per
well in the presence of 1.5 mm butyric acid two days prior to
experiments. BHK21-tTA cells transfected with antisense
CHC were grown in DMEM supplemented with 10% (v ⁄ v)
fetal bovine serum, 100 units Æ mL
)1
penicillin, 100 lgÆmL
)1
streptomycin, 0.5 lgÆmL
)1
puromycin, 0.2 mgÆmL
)1
geneti-
cin and 2 lgÆmL
)1
tetracycline. Two days before experiments
the cells were seeded with and without tetracycline in 24-well
plates at a density of 2 · 10
4
cells per well.
Endocytosis of TAG- and biotin-labeled Stx, CT
and transferrin
Endocytosis of Stx, CT and transferrin was performed
essentially as described previously [8,30]. Briefly, this special
method is based on double-labeling of the proteins of inter-
est with both a capture-label (biotin) and a detection-label
(electrochemiluminescent label), which are selectively cap-
tured and quantitated in a highly specialized electrochemi-
luminescent detection instrument produced by BioVeris

Corporation (Gaithersburgh, MD, USA). Stx, CT or
transferrin were labeled with the detection label BV-TAGÒ
(BioVeris Corporation), which was stably bound to the pro-
teins via an amide bond. The central atom of this specific
label is a tris (bipyridine)-chelated ruthenium (II), which
emits light when electrochemically stimulated, and the pho-
tons are measured in a photomultiplier tube in less than a
second. Capture of the protein is mediated via streptavidin-
coated magnetic beads, and the protein must therefore be
simultaneously labeled with biotin. Once bound to the cell
surface, neither Stx nor CT is easily removed, even with
proteases. Therefore, to distinguish internalized toxin from
total cell-associated toxin (bound + internalized), the
toxins were labeled with the reducible EZ-Link Sulfo-
NHS-SS-Biotin (Pierce, Rockford, IL, USA). Then, upon
internalization of labeled toxin, the biotin linked to surface-
bound toxin was cleaved off in some of the wells by the
cell-impermeable reducing agent MESNa, thereby prevent-
ing detection of the still surface-bound toxins [38].
In the experiments, the cells were washed once with
Hepes buffered MEM and then incubated with TAG- and
biotin-labeled Stx (0.33 nm), CT (0.12 nm) or transferrin
(0.64 nm) in Hepes buffered MEM in the presence of BSA
(0.2%) for different time periods at 37 °C. In some experi-
ments the cells were preincubated with the labeled toxins
for 30 min at 0 °C, followed by a brief wash, and endocyto-
sis was measured after incubation at 37 °C for the indicated
time periods. After internalization of toxin, the cells were
washed with cold buffer (0.14 m NaCl, 2 mm CaCl
2

,20mm
Hepes; pH 8.6), and one half of the wells were treated with
0.1 m MESNa in the same buffer for 20 min at 0 °Cto
reduce the SS-linked biotin in cell surface-bound toxin,
while the other half were mock treated. The cells were
washed in cold buffer (0.14 m NaCl, 2 mm CaCl
2
,20mm
Hepes; pH 7.0) and lysed (1% Triton, 60 mm n-octyl b-d-
glucopyranoside, 100 mm NaCl, 5 mm MgCl
2
,50mm
Hepes). Streptavidin-coated magnetic beads (Dynal, Oslo,
Norway) were added to the lysate, and the samples were
shaken for 20 min. The amount of TAG-labeled toxin cap-
tured by the beads was detected in a BioVeris Detection
System instrument (BioVeris Corporation). Counts from
cells treated with MESNa represent the amount of internal-
ized toxin, while counts from untreated cells represent
the total amount of toxin associated with the cells (bound
+ internalized). Endocytosis of Stx was reported as inter-
nalized toxin in percent of total cell-associated toxin. The
background in this system, as measured by treating the cells
with unlabeled toxin alone, was low ( 1200 counts) and
equal to counting lysis buffer alone. The background value
was subtracted from all the values in each experiment. The
endocytic values, with background subtracted, ranged from
5000 to 10 000, depending on the cell line, and the values
for total cell-associated toxin ranged from 20 000 to
40 000. It should be noted that even for the lowest experi-

mental values used, the signal : background ratio was not
lower than  3. Furthermore, upon binding of labeled toxin
to cells on ice, treatment with MESNa reduced the detected
amount of surface-bound toxin to 98%. Thus, there was a
2% constant background value after MESNa treatment,
which was subtracted from all the experimental values.
Stx produced by S. dysenteriae and Stx1 produced by
E. coli are virtually identical and differ only in one amino
acid. Replacing Stx with Stx1 in the endocytosis experiment
Endocytosis of Shiga toxin M. L. Torgersen et al.
4110 FEBS Journal 272 (2005) 4103–4113 ª 2005 FEBS
shown in Fig. 1, gave nearly identical results (data not
shown). Unless otherwise stated, Stx was substituted with
the commercially available Stx1 in the remaining experi-
ments. Similarly, identical results were obtained by StxB-
Sulf
2
and StxB from S. dysenteriae (data not shown), and
both proteins have been used.
Preparation of StxB
The StxB containing a tandem of sulfation sites in the
C-terminus (StxB-Sulf
2
) was produced in E. coli BL21
(DE3) cells essentially as described previously [39]. Briefly,
a 10 mL overnight bacterial culture grown at 37 °C was
inoculated in 500 mL of LB medium and further grown to
an attenuance at 600 nm of 0.6. The culture was heat-
induced for 4 h at 42 °C, and the cells were harvested by
centrifugation. The pellet was washed twice with 10 mm

Tris ⁄ HCl (pH 8.0), resuspended in 25% (w ⁄ v) sucrose,
1mm Na
2
EDTA, and 10 mm Tris ⁄ HCl (pH 8.0), and
gently shaken at 30 °C for 10 min. Cells were harvested by
centrifugation and resuspended in ice-cold distilled water.
After centrifugation the supernatant was dialyzed overnight
against 20 mm Tris ⁄ HCl (pH 7.5), loaded on a Resource Q
column (Amersham Biosciences, Uppsala, Sweden), and
eluted with a 0–600 mm NaCl gradient in 20 mm Tris ⁄ HCl
(pH 7.5). At physiological pH StxB is in a stable pentameric
form, and the dialysis was actually performed in tubing with
cut-off 12–14 000 Da. Purified StxB is therefore structurally
equivalent to the B-moiety of the holotoxin [40–42].
Immunofluorescence
HeLa cells grown on coverslips were washed with cold
Hepes medium and cooled to 10 °C before incubation with
Alexa-labeled Stx (1.3 nm) and Cy-5 labeled transferrin
(130 nm) for 40 min at 10 °C. The cells were washed twice
with NaCl ⁄ P
i
and fixed with 3% (v ⁄ v) paraformaldehyde in
NaCl ⁄ P
i
, before analysis in a LSM 510 Meta confocal
microscope (Zeiss, Oberkochen, Germany). Pictures were
taken of thin single plane sections. The extent of colocaliza-
tion between Stx (green channel) and transferrin (blue
channel, changed to red for quantification) was quantified
as reported previously [8] by calculating the ratio between

the number of yellow (¼ colocalization) pixels (fluorescence
level between 150 and 255) and the number of green pixels
(fluorescence level between 150 and 255) using the Adobe
photoshop 7.0 software.
Estimation of cell surface-bound toxin
concentration
To be able to compare the local concentration of Stx on
the cell surface and in the medium, the labeled toxin bound
to the cells were assumed to be distributed in a restricted
volume surrounding the cell from the outside of the plasma
membrane and 10 nm outwards in the medium. Counts per
lL in this volume were compared to the counts of free
toxin molecules per lL of medium. The volume of a HEp-2
cell has been estimated to be 4.5 · 10
)12
lL, and the num-
ber of cells after one day of growth was approximated to
be 7.5 · 10
4
per well.
Acknowledgements
We thank Dr Tore-Geir Iversen for critically reading
the manuscript. This work was supported by The Nor-
wegian Cancer Society, The Norwegian Research
Council for Science and the Humanities, the Novo-
Nordisk Foundation, the Jahre Foundation and Jea-
nette and Søren Bothners legacy.
References
1 Acheson DW, Kane AV & Keusch GT (2000) Shiga
toxins. Methods Mol Biol 145, 41–63.

2 Sandvig K, Grimmer S, Lauvrak U, Torgersen L,
Skretting G, van Deurs B & Iversen G (2002) Pathways
followed by ricin and Shiga toxin into cells. Histochem
Cell Biol 117, 131–141.
3 Sandvig K & van Deurs B (2002) Transport of protein
toxins into cells: pathways used by ricin, cholera toxin
and Shiga toxin. FEBS Lett 529, 49–53.
4 Sandvig K (2001) Shiga toxins. Toxicon 39, 1629–1635.
5 Karmali MA (2004) Prospects for preventing serious
systemic toxemic complications of Shiga toxin-produ-
cing Escherichia coli infections using Shiga toxin recep-
tor analogues. J Infect Dis 189, 355–359.
6 Khine AA & Lingwood CA (1994) Capping and recep-
tor-mediated endocytosis of cell-bound verotoxin
(Shiga-like toxin). 1: Chemical identification of an
amino acid in the B-subunit necessary for efficient
receptor glycolipid binding and cellular internalization.
J Cell Physiol 161, 319–332.
7 Nichols BJ, Kenworthy AK, Polishchuk RS, Lodge R,
Roberts TH, Hirschberg K, Phair RD & Lippincott-
Schwartz J (2001) Rapid Cycling of Lipid Raft Markers
between the Cell Surface and Golgi Complex. J Cell
Biol 153, 529–542.
8 Lauvrak SU, Torgersen ML & Sandvig K (2004) Effi-
cient endosome-to-Golgi transport of Shiga toxin is
dependent on dynamin and clathrin. J Cell Sci 117,
2321–2331.
9 Johannes L, Tenza D, Antony C & Goud B (1997) Ret-
rograde transport of KDEL-bearing B-fragment of
Shiga toxin. J Biol Chem 272, 19554–19561.

10 Kim JH, Johannes L, Goud B, Antony C, Lingwood
CA, Daneman R & Grinstein S (1998) Noninvasive
measurement of the pH of the endoplasmic reticulum at
M. L. Torgersen et al. Endocytosis of Shiga toxin
FEBS Journal 272 (2005) 4103–4113 ª 2005 FEBS 4111
rest and during calcium release. Proc Natl Acad Sci
USA 95, 2997–3002.
11 Mallard F, Antony C, Tenza D, Salamero J, Goud B &
Johannes L (1998) Direct pathway from early ⁄ recycling
endosomes to the Golgi apparatus revealed through the
study of shiga toxin B-fragment transport. J Cell Biol
143, 973–990.
12 Girod A, Storrie B, Simpson JC, Johannes L, Goud B,
Roberts LM, Lord JM, Nilsson T & Pepperkok R
(1999) Evidence for a COP-I-independent transport
route from the Golgi complex to the endoplasmic reti-
culum. Nat Cell Biol 1, 423–430.
13 White J, Johannes L, Mallard F, Girod A, Grill S,
Reinsch S, Keller P, Tzschaschel B, Echard A, Goud B
& Stelzer EH (1999) Rab6 coordinates a novel Golgi to
ER retrograde transport pathway in live cells. J Cell
Biol 147, 743–760.
14 Wilcke M, Johannes L, Galli T, Mayau V, Goud B &
Salamero J (2000) Rab11 regulates the compartmentali-
zation of early endosomes required for efficient trans-
port from early endosomes to the trans-golgi network.
J Cell Biol 151, 1207–1220.
15 Falguieres T, Mallard F, Baron C, Hanau D, Lingwood
C, Goud B, Salamero J & Johannes L (2001) Targeting
of Shiga toxin B-subunit to retrograde transport route

in association with detergent-resistant membranes. Mol
Biol Cell 12, 2453–2468.
16 Johannes L, Pezo V, Mallard F, Tenza D, Wiltz A,
Saint-Pol A, Helft J, Antony C & Benaroch P (2003)
Effects of HIV-1 Nef on retrograde transport from the
plasma membrane to the endoplasmic reticulum. Traffic
4, 323–332.
17 Saint-Pol A, Yelamos B, Amessou M, Mills IG, Dugast
M, Tenza D, Schu P, Antony C, McMahon HT,
Lamaze C & Johannes L (2004) Clathrin Adaptor
epsinR Is Required for Retrograde Sorting on Early
Endosomal Membranes. Dev Cell 6, 525–538.
18 Khine AA, Tam P, Nutikka A & Lingwood CA (2004)
Brefeldin A and filipin distinguish two globotriaosyl
ceramide ⁄ verotoxin-1 intracellular trafficking pathways
involved in Vero cell cytotoxicity. Glycobiology 14, 701–
712.
19 Tai G, Lu L, Wang TL, Tang BL, Goud B, Johannes L
& Hong W (2004) Participation of the syntaxin
5 ⁄ Ykt6 ⁄ GS28 ⁄ GS15 SNARE complex in transport from
the early ⁄ recycling endosome to the trans-Golgi net-
work. Mol Biol Cell 15, 4011–4022.
20 Lu L, Tai G & Hong W (2004) Autoantigen Golgin-97,
an effector of Arl1 GTPase, participates in traffic from
the endosome to the trans-golgi network. Mol Biol Cell
15, 4426–4443.
21 Martinez-Alonso E, Egea G, Ballesta J & Martinez-
Menarguez JA (2005) Structure and dynamics of the
Golgi complex at 15 degrees C: low temperature induces
the formation of Golgi-derived tubules. Traffic 6, 32–44.

22 Sandvig K, Olsnes S, Brown JE, Petersen OW & van
Deurs B (1989) Endocytosis from coated pits of Shiga
toxin: a glycolipid-binding protein from Shigella dysen-
teriae 1. J Cell Biol 108, 1331–1343.
23 Sandvig K, Prydz K, Ryd M & van Deurs B (1991)
Endocytosis and intracellular transport of the glyco-
lipid-binding ligand Shiga toxin in polarized MDCK
cells. J Cell Biol 113, 553–562.
24 Sandvig K, Dubinina E, Garred O, Prydz K, Kozlov
JV, Van Hansen SH, Skretting G, Llorente A, Nicoziani
P, van Deurs B & Sandvig K (2001) Endosome to Golgi
transport of ricin is independent of clathrin and of the
Rab9- and Rab11-GTPases. Mol Biol Cell 12, 2099–
2107.
25 Iversen, T-G, Skretting, G, Llorente, A, Nicoziani, P,
van Deurs, B, & Sandvig, K (2001) Endosome to Golgi
transport of ricin is independent of clathrin and of the
Rab9- and Rab11-GTPases. Mol. Biol. Cell 12, 2099–
2107.
26 Llorente A, Prydz K, Sprangers M, Skretting G, Kolset
SO & Sandvig K (2001) Proteoglycan synthesis is
increased in cells with impaired clathrin-dependent
endocytosis. J Cell Sci 114, 335–343.
27 Iversen TG, Skretting G, van Deurs B & Sandvig K
(2003) From the Cover: Clathrin-coated pits with long,
dynamin-wrapped necks upon expression of a clathrin
antisense RNA. Proc Natl Acad Sci USA 100, 5175–5180.
28 Shimizu T, Hamabata T, Yoshiki A, Hori T, Ito S,
Takeda Y & Hayashi H (2003) An association of 27-
and 40-kDa molecules with glycolipids that bind A-B

bacterial enterotoxins to cultured cells. Biochim Biophys
Acta 1612, 186–194.
29 Puri C, Tosoni D, Comai R, Rabellino A, Segat D,
Caneva F, Luzzi P, Di Fiore PP & Tacchetti C (2005)
Relationships between EGFR Signaling-competent and
Endocytosis-competent Membrane Microdomains. Mol
Biol Cell 16, 2704–2718.
30 Torgersen ML, Skretting G, van Deurs B & Sandvig K
(2001) Internalization of cholera toxin by different
endocytic mechanisms. J Cell Sci 114, 3737–3747.
31 Jackson MP, Newland JW, Holmes RK & O’Brien AD
(1987) Nucleotide sequence analysis of the structural
genes for Shiga-like toxin I encoded by bacteriophage
933J from Escherichia coli. Microb Pathog 2, 147–153.
32 Sunyach C, Jen A, Deng J, Fitzgerald KT, Frobert Y,
Grassi J, McCaffrey MW & Morris R (2003) The
mechanism of internalization of glycosylphosphatidyl-
inositol-anchored prion protein. EMBO J 22, 3591–
3601.
33 Fjeldstad K, Pedersen ME, Vuong TT, Kolset SO,
Nordstrand LM & Prydz K (2002) Sulfation in the
Golgi lumen of Madin–Darby canine kidney cells is
inhibited by brefeldin A and depends on a factor pre-
sent in the cytoplasm and on Golgi membranes. J Biol
Chem 277, 36272–36279.
Endocytosis of Shiga toxin M. L. Torgersen et al.
4112 FEBS Journal 272 (2005) 4103–4113 ª 2005 FEBS
34 Guibinga GH, Miyanohara A, Esko JD & Friedmann T
(2002) Cell surface heparan sulfate is a receptor for
attachment of envelope protein-free retrovirus-like parti-

cles and VSV-G pseudotyped MLV-derived retrovirus
vectors to target cells. Mol Ther 5, 538–546.
35 Sandvig K & van Deurs B (1996) Endocytosis, intracel-
lular transport, and cytotoxic action of Shiga toxin and
ricin. Physiol Rev 76, 949–966.
36 Marsh M & McMahon HT (1999) The structural era of
endocytosis. Science 285, 215–220.
37 Fraser ME, Fujinaga M, Cherney MM, Melton-Celsa
AR, Twiddy EM, O’Brien AD & James MN (2004)
Structure of shiga toxin type 2 (Stx2) from Escherichia
coli O157:H7. J Biol Chem 279, 27511–27517.
38 Smythe E, Redelmeier TE & Schmid SL (1992) Recep-
tor-mediated endocytosis in semiintact cells. Meth
Enzymol 219, 223–234.
39 Su GF, Brahmbhatt HN, Wehland J, Rohde M &
Timmis KN (1992) Construction of stable LamB-Shiga
toxin B subunit hybrids: analysis of expression in Sal-
monella typhimurium aroA strains and stimulation of B
subunit-specific mucosal and serum antibody responses.
Infect Immun 60, 3345–3359.
40 Pina DG & Johannes L (2005) Cholera and Shiga toxin
B-subunits: thermodynamic and structural considera-
tions for function and biomedical applications. Toxicon
45, 389–393.
41 Pina DG, Gomez J, Villar E, Johannes L & Shnyrov
VL (2003) Thermodynamic analysis of the structural
stability of the shiga toxin B-subunit. Biochemistry 42,
9498–9506.
42 Saleh MT & Gariepy J (1993) Local conformational
change in the B-subunit of Shiga-like toxin 1 at endo-

somal pH. Biochemistry 32, 918–922.
M. L. Torgersen et al. Endocytosis of Shiga toxin
FEBS Journal 272 (2005) 4103–4113 ª 2005 FEBS 4113