Báo cáo khoa học: A pH-dependent conformational change in EspA, a component of the Escherichia coli O157:H7 type III secretion system potx

A pH-dependent conformational change in EspA, a
component of the Escherichia coli O157:H7 type III
secretion system
Tomoaki Kato
1,2
, Daizo Hamada
2
, Takashi Fukui
2
, Makoto Hayashi
1
, Takeshi Honda
3
,
Yoshikatsu Murooka
1
and Itaru Yanagihara
2
1 Department of Biotechnology, Graduate School of Engineering, Osaka University, Japan
2 Department of Developmental Infectious Diseases, Research Institute, Osaka Medical Center for Maternal and Child Health, Japan
3 Department of Bacterial Infections, Research Institute for Microbial Diseases, Osaka University, Japan
Enterohaemorrhagic and enteropathogenic Escherichia
coli (EHEC and EPEC, respectively) cause outbreaks
of serious diarrhoea. These bacteria express type III
secretion systems [1], which consist of various protein
components encoded at the locus of enterocyte efface-
ment, LEE [2–5]. To date, type III secretion systems
have been identified in more than 20 pathogenic bac-
terial species [6]. The type III secretion system is a fila-
mentous multiprotein complex that assembles across
the bacterial and host cell surfaces. For EHEC and

EPEC, such complex structures, which include the pro-
teins, EspA, EspB, EspD [7,8], probably permit direct
delivery of effector proteins, such as, Tir [9–11], EspF
[12,13], EspG [14] and Orf19 [15], into the host cell
[16].
EspA is a major component of this large, transiently
expressed, filamentous surface organelle [17,18]. EspA
oligomerization may be mediated by interactions
between coiled-coil regions of individual EspA mole-
cules [19] in a manner similar to that of falgellin
Keywords
ANS binding; CD; FT-IR; partially unfolded;
sedimentation equilibrium
Correspondence
D. Hamada, Department of Developmental
Infectious Diseases, Research Institute,
Osaka Medical Center for Maternal and
Child Health, 840 Murodo, Izumi,
Osaka 594-1011, Japan
Fax: +81 725 57 3021
Tel: +81 725 56 1220
E-mail:
(Received 21 September 2004, revised
1 March 2005, accepted 1 April 2005)
doi:10.1111/j.1742-4658.2005.04697.x
pH-Dependent structural changes for Escherichia coli O157:H7 EspA were
characterized by CD, 8-anilino-2-naphthyl sulfonic acid (ANS) fluores-
cence, and sedimentation equilibrium ultracentrifugation. Far- and near-
UV CD spectra, recorded between pH 2.0 and 7.0, indicate that the protein
has significant amounts of secondary and tertiary structures. An increase in

ANS fluorescence intensity (in the presence of EspA) was observed at aci-
dic pH; whereas, no increased ANS fluorescence was observed at pH 7.0.
These results suggest the presence of a partially unfolded state. Interest-
ingly, urea-induced unfolding transitions, monitored by far-UV CD spectro-
scopy, showed that the protein is destabilized at pH 2.0 as compared with
EspA at neutral pH. Although increased ANS fluorescence was observed
at pH 3.0, the urea-induced unfolding curve is similar to that found at
pH 7.0. This result suggests the presence, at pH 3.0, of an ordered, but
partially unfolded state, which differs from typical molten globule. The
results of analytical ultracentrifugation and infrared spectroscopy indicate
that EspA molecules associate at pH 7.0, suggesting the formation of short
filamentous oligomers containing a-helical structures, whereas the protein
tend to form nonspecific aggregates containing intermolecular b-sheets at
pH 2.0. Our experiments indicate that EspA has the potential to spontane-
ously form filamentous oligomers at neutral pH; whereas the protein is
partially unfolded, assuming different conformations, at acidic pH.
Abbreviations
ANS, 8-anilinonaphthalene-1-sulfonic acid; EHEC, enterohaemorrhagic Escherichia coli; EPEC, enteropathogenic Escherichia coli; FT-IR,
Fourier transform infrared; LB, Luria–Bertani.
FEBS Journal 272 (2005) 2773–2783 ª 2005 FEBS 2773
molecules, which assemble to form flagella filaments
[20]. The EspA-containing filamentous apparatus may
form a conduit for translocation of bacterial proteins
into host cells [21]. Recently, a model of EspA fila-
ments has been built based on negative-stain EspA
electron micrographs [18]. Interestingly, the model is a
helical tube with a diameter of 120 A
˚
, enclosing a cen-
tral channel of 25 A

˚
diameter, and has an axial rise of
4.6 A
˚
per subunit. EspA filaments may attach to host
cells via an EspB ⁄ D pore-forming complex [22] and
the EspB ⁄ D complex may also specifically interact with
the host-target protein, a-catenin [23]. Such a super-
structure, formed by EspA⁄ B ⁄ D and a-catenin, facili-
tates the delivery of effector proteins into host cells
[16].
Although there is information available concerning
the roles and structural properties of EspA filaments,
to date, the conformation and the thermodynamic
properties of EspA have not been characterized.
In this study, we characterized certain conformational
and thermodynamic properties of EspA in solution
using spectroscopic and physicochemical techniques.
Far-UV CD shows that the protein has a substantial
amount of secondary structure throughout the pH
range of 2.0–7.0. However, an analysis of 8-anilino-
2-naphthyl sulfonic acid (ANS) fluorescence (in the
presence of EspA) suggests that a conformational
transition occurs between pH 3.0 and 5.0, with expo-
sure of hydrophobic protein surfaces. Consistent with
this observation, urea-induced EspA unfolding transi-
tions, as followed by far-UV CD, indicate that the
folded structure is less stable at pH 2.0. A sedimenta-
tion equilibrium study shows that EspA forms oligo-
mers at pH 7.0, indicating an ability by EspA to form

filamentous structures. The data now reported suggest
that EspA, at near physiological conditions, assumes
short filamentous oligomers, but dissociates into parti-
ally unfolded species at acidic pH.
Results
CD
The secondary structure prediction for EspA, based on
its amino acid sequence, suggests that the protein is
predominantly a-helical in conformation (Fig. 1). The
secondary structures of the recombinant EspA pre-
pared here were analysed by far-UV CD spectra. In
this study, the recombinant EspA protein was prepared
under either native or denaturing conditions. To clarify
whether both preparations yielded protein with similar
propertiees, we first compared the CD spectra of EspA
prepared under the different conditions.
Recombinant EspA prepared from soluble fractions
without unfolding the protein, at pH 7.0 and 20 °C,
showed a CD spectrum typical of a protein with a
significant amount of secondary structure (Fig. 2A).
Importantly, the spectrum for the sample prepared by
urea-solubilized cells was almost completely super-
imposable on the above spectrum. This observation
Fig. 1. Secondary structure prediction for EspA based on its amino
acid sequence. H and E refer to a-helical and b-strand propensities,
respectively.
Fig. 2. CD spectra of EspA at various pH values. (A) Far- and (B)
near-UV CD spectra at pH 2.0 (broken line), pH 3.0 (dotted line),
and pH 7.0 (continuous line). Circles indicate the spectrum at
pH 7.0 for recombinant EspA prepared from soluble fraction of cell

lysate. (C) Plot of ellipticity at 222 nm vs. pH.
pH-Dependent EspA conformational change T. Kato et al.
2774 FEBS Journal 272 (2005) 2773–2783 ª 2005 FEBS
suggests that our preparation of recombinant EspA
using urea, which included unfolding and refolding
steps was successful, and that this protein reversibly
unfolds and refolds, at least under the controlled con-
ditions used here.
The ellipticity at 222 nm for EspA at pH 7.0 was
)13 600 degÆcm
2
Ædmol
)1
. This ellipticity value yields an
a-helical content of 37.2% when used in the equation:
f
H
¼ Àð½h
222
þ 2340Þ=30300
where f
H
and [h]
222
are the a-helical fraction and the
ellipticity at 222 nm, respectively [24]. This value,
derived from the pH 7.0 CD spectrum, is smaller than
that estimated from the secondary structure prediction
using the amino acid sequence (62.0% and 8.3% for
a-helices and b-sheets, respectively, Table 1). The sec-

ondary structure contents, estimated by the program
CDPro, are 39.6% and 13.6% for a-helices and
b-sheets, respectively. This a-helical value is also smal-
ler than that predicted using the amino acid sequence
(62.0% and 8.3% for a-helices and b-sheets, respect-
ively, Table 1).
Although the intensity is significantly low, the near-
UV CD spectrum of EspA at pH 7.0 and 20 °C,
showed a minimum and a maximum around 280 and
290 nm. As in the case of far-UV CD, the near-UV
CD spectrum at pH 2.0 was similar to the spectrum at
pH 7.0, although the intensity of each peak bacome
slightly smaller. It is in the near-UV region that aro-
matic residues display optical activity. EspA contains
five tyrosines at positions 22, 51, 53, 110 and 182, and
no tryptophans. Therefore, the shape of the near-UV
CD spectra of EspA suggests the presence of some ter-
tiary contacts around at least one of the tyrosines both
at pH 7.0 and 2.0 (Fig. 2B).
To gain further insight into the conformational
properties of EspA, we recorded far-UV CD spectra
for solutions with the pH adjusted between 2.0 and
7.0. Interestingly, when the solution pH was between
3.0 and 7.0, the spectra are almost identical and the
derived secondary structure estimates are similar to
each other (Fig. 2C and Table 1). The pH 2.0 spec-
trum also indicates a significant amount of secondary
structure, although the spectral intensity is smaller
than those obtained at higher pH. This observation
suggests that at pH 2.0 EspA is less ordered than at

pH 3.0–7.0.
ANS binding
ANS binds to solvent-accessible hydrophobic surfaces
and when bound its fluorescence intensity at % 500 nm
increases. This property of ANS is often used to detect
partially unfolded protein intermediates [25], e.g. mol-
ten globules, which are compact intermediates with
significant amounts of native-like secondary structure,
but with disordered tertiary contacts and solvent-
exposed hydrophobic clusters [26–34]. To determine if
partially unfolded EspA species are present as a result
of solution conditions, we recorded ANS fluorescence
spectra, with EspA present at various pH conditions.
Between pH 6.0 and 9.0 the ANS fluorescence
was insignificant (Fig. 3), suggesting that negligible
amounts of hydrophobic surfaces are solvent-accessible
However, ANS fluorescence increased when the pH
decreased from 6.0 to 2.0 (Fig. 3). This observation
suggests that hydrophobic surfaces become exposed
upon decreasing the pH. Since the protein maintains a
significant amount of secondary structure (as estimated
Table 1. EspA Secondary structure composition at various pH val-
ues as estimated using the far-UV CD spectral data. Values were
calculated using CDPro [44,45].
Conditions a-Helix (%) b-Sheet (%) Turn (%) Others (%)
pH 2.0 32.6 ± 3.0 16.1 ± 2.4 21.9 ± 0.9 29.9 ± 0.4
pH 3.0 40.9 ± 6.9 12.0 ± 4.1 19.6 ± 2.1 29.1 ± 2.0
pH 5.0 39.8 ± 8.1 13.0 ± 1.1 19.0 ± 2.0 28.5 ± 1.0
pH 7.0 39.6 ± 7.1 13.6 ± 5.2 19.7 ± 1.7 27.6 ± 0.4
Predicted values

a
62.0 8.3 29.7
b
a
Estimated from the secondary structure prediction (Fig. 1) using
the
PHDSEC algorithm available at the PREDICTPROTEIN server [46–48].
b
The value is for non-a-helical and non-b-strand regions.
Fig. 3. ANS fluorescence at 460 nm as a function of pH. Circles
indicate the raw data. The line is drawn only to assist the reader
and has no theoretical relevance. The approximated baselines for
N
II
and N
I
(see text in detail) are shown by dotted and broken lines,
respectively.
T. Kato et al. pH-Dependent EspA conformational change
FEBS Journal 272 (2005) 2773–2783 ª 2005 FEBS 2775
from far-UV CD spectra between pH 2.0 and 5.0,
Fig. 2), the results of the ANS study suggest formation
of a partially unfolded state, possibly similar to the
a-lactalbumin molten globule characterized at acidic
pH [34].
Sedimentation equilibrium
Under physiological conditions, during EHEC or
EPEC infection, EspA is associated with filamentous
structures. We therefore tested, using sedimentation
equilibrium ultracentrifugation, whether recombinant

EspA has the potential to form oligomers.
Figure 4 shows the results for the EspA sedimenta-
tion equilibrium experiments at pH 7.0 and 20 °C. If a
protein solution contains only a single molecular
weight species, then a plot of the natural logarithm of
the protein absorption at 280 nm [ln(A
280
)] vs. the
square of the radial distance (r
2
) shows a linear
correlation between ln(A
280
) and r
2
. However, the data
(Fig. 4A) indicate that ln(A
280
) exponentially increases
with increased r
2
. This sedimentation equilibrium pro-
file indicates either the presence of large protein oligo-
mers or a contribution to the plot by nonideal solution
behaviour. Probably, the solution behaves as an ideal
system under the experimental conditions, i.e., 10 mm
sodium phosphate, pH 7.0, 100 mm NaCl. Thus, it is
unlikely that the curvature shown in Fig. 4A is caused
by nonideal behaviour.
Figure 4B shows that M

app
increases with an
increase in the concentration of EspA. The data of
Fig. 4B suggest that the size distribution of EspA ran-
ges from that of the monomer to approximately that
of a 30-mer when the protein concentration is
1mgÆmL
)1
, i.e. 44 lm.
We also attempted to analyse the sedimentation pro-
file of the partially unfolded state that exists at pH 2.0.
However, it was extremely difficult to obtain the exact
size of protein at pH 2.0 probably due to the forma-
tion of irreversible aggregates during the long period
of centrifugation. Although no visible precipitates were
found at the beginning, protein absorption started to
decrease after about 24 h, and become almost unde-
tectable after 48 h. This may be caused by the require-
ment of high protein concentration (> 1 mgÆmL
)1
) due
to the lack of tryptophan residues in EspA for reliable
detection as well as the need for a long equilibration
period (> 24 h) essential for the sedimentation equilib-
rium study. The result is, however, consistent with the
idea that the protein is partially unfolded at pH 2.0,
because partially unfolded species are generally prone
to form nonspecific aggregates (see below). Thus, com-
pared with other simple spectroscopic measurements
such as CD, ultracentrifugation may generally not be

suitable for the analysis of partially folded proteins
which are prone to aggregate.
FT-IR spectroscopy
The previous sedimentation equilibrium study indica-
ted that after a long incubation at pH 2.0 the EspA
solution contains aggregates, although these are invis-
ible just after preparation of the sample. FT-IR spectro-
scopy also confirmed the presence of molecular
species containing intermolecular b-strands typical for
nonspecific aggregates.
As indicated by CD, the soluble EspA at pH 7.0
shows an IR spectrum suggestive of the formation of
a-helical structures with a peak around 1650 cm
)1
(Fig. 5). However, the spectrum taken for the solution
at pH 2.0 has an additional maximum peak around
1620 cm
)1
which is characteristic for the intermole-
Fig. 4. Sedimentation equilibrium. (A) Plot of the logarithm of the
absorbance at 280 nm, A
280
, as a function of the square of the
radial distance, r
2
. Data were collected at pH 7.0 with 1.0 mgÆmL
)1
EspA. (B) Plot of M
app
vs. protein concentration.

pH-Dependent EspA conformational change T. Kato et al.
2776 FEBS Journal 272 (2005) 2773–2783 ª 2005 FEBS
cular b-sheets usually formed in the nonspecific aggre-
gates.
Importantly, precipitates are also formed at pH 7.0
in the presence of EspA > 1 mgÆmL
)1
. The IR spec-
trum for these aggregates, however, is significantly
similar to the spectrum taken for the soluble EspA
(Fig. 5). This observation suggests that EspA has an
intrinsic potential to self-associate into oligomeric
structures, which consist of a-helical secondary struc-
tures.
Urea-induced unfolding
The stability of EspA at various pH values was ana-
lysed using far-UV CD spectroscopy. By plotting the
ellipticity at 222 nm as a function of urea concentra-
tion, cooperative unfolding transitions were obtained
at all pH values (Fig. 6). Between pH 3.0 and 7.0, the
transitions occurred between 3.0 and 6.0 m urea. How-
ever, the transition region shifted towards lower urea
concentrations of about 0.0–3.0 m at pH 2.0. This
observation is qualitatively consistent with the ANS
binding results, which show that the protein, at
pH 2.0, assumes a partially unfolded conformation
with exposed hydrophobic surfaces. Interestingly, the
stability of the protein at pH 3.0 seems comparable to
that at pH 7.0. This observation would seem to be
inconsistent with the pH 3.0 ANS binding experiment

as that experiment indicates a degree of unfolding
resulting in hydrophobic surface solvent-exposure.
Therefore, between pH 3 and 5, a partially unfolded
state with highly ordered native-like tertiary contacts,
but also with fluctuating regions, may exist. This
conformational state is clearly distinguishable from the
typical molten globule formed by EspA at pH 2.0.
Discussion
pH-dependence of EspA conformations
The present analysis suggests that the amount of EspA
secondary structure, at various pH conditions, is
highly conserved, even at pH 2.0. However, ANS bind-
ing experiments indicate that a conformational change
occurred upon decreasing pH. The characteristics of
this conformational change are consistent with the for-
mation of a partially unfolded species, probably sim-
ilar to a molten globule. Molten globules are compact
denatured states with significant amounts of native-like
secondary structure, but with disrupted tertiary inter-
actions [26–34]. Although peak intensity is slightly dif-
ferent, the near-UV CD spectrum at pH 7.0 is closely
similar to that at pH 2.0 (Fig. 2B). This is apparently
inconsistent with the idea that the conformational spe-
cies of EspA at pH 2.0 is in a typical molten globule
state. In this sense, the partially unfolded structure at
pH 2.0, which exposes hydrophobic clusters to the sol-
vent, may contain rather rigid tertiary conformation
compared with the classical molten globule state. It
should be noted that the urea-induced denaturation
data indicated a decreased stability and cooperativity

against urea-induced unfolding for EspA at pH 2.0
compared with that at pH 7.0. This suggests that
some conformational transitions may occur around
Fig. 6. Urea-induced EspA unfolding transitions at various pH val-
ues, 20 °C. The transition curves are obtained from far-UV CD
spectra at pH 2.0 (triangles), pH 3.0 (squares) and pH 7.0 (circles).
The approximated baselines for folded (N
I
or N
II
) and unfolded
states are drawn by dotted and broken lines, respectively, The ideal
ellipticity for 50% of folded or unfolded species is shown by a thin
line.
Fig. 5. FT-IR spectroscopy of EspA. The spectra at pH 2.0 (broken
line), aggregates formed at pH 7.0 (dotted line), and soluble fraction
at pH 7.0 (continuous line).
T. Kato et al. pH-Dependent EspA conformational change
FEBS Journal 272 (2005) 2773–2783 ª 2005 FEBS 2777
pH 2.0–3.0. Importantly, at pH 2.0, the ellipticity at
222 nm decreased compared with the value at pH 3.0–
7.0. Thus, some of the a-helical structure formed at
pH 3.0–7.0 may be disrupted at pH 2.0, whereas ter-
tiary contacts, at least, around one of the tyrosine resi-
dues are conserved.
Recently, the three-dimensional structure of EspA
complexed with its chaperone, CesA, has been solved
by X-ray crystallography [35]. In this model, only the
N-terminal 29 and C-terminal 43 residues (amino acid
positions at 31–59 and 148–190) of EspA correspond-

ing to the binding interface of CesA could be clearly
solved. The other regions corresponding to the amino
acid positions between 60 and 147 could not be solved,
possibly due to the conformational disorder or mul-
tiple conformations. If the unsolved regions in the
EspA–CesA complex structure are disordered, the
a-helical content of EspA should be 37.5%. This value
is highly consistent with the a-helical content estimated
here from far-UV CD spectra of EspA at pH 7.0
(39.6%). It is generally considered that the native-like
secondary structures are present in the partially folded
state of a protein. Thus, it would be natural to assume
that the two a-helices of EspA shown in the EspA–
CesA complex structures may be also formed in the
partially folded state of EspA at pH 2.0. According to
the EspA–CesA complex structure, only Y53 forms
tertiary contacts with the C-terminal a-helix of EspA
and CesA, and other tyrosines located in these a-heli-
ces are exposed to the solvent. Therefore, the near-UV
CD signals observed at pH 7.0 and 2.0 in Fig. 2B
might be responsible for the formation of tertiary con-
tacts around Y53. The formation of nonspecific aggre-
gates which occurred at pH 2.0 in the presence of high
concentration of EspA indicate that the oligomeric
EspA at pH 7.0 can tend to dissociate into monomers
at pH 2.0 since the oligomerization into native struc-
ture should prevent the formation of nonspecific aggre-
gates. In this sense, the near-UV CD signal observed
at pH 2.0 can be responsible for the intramolecular
tertiary contacts around Y53, whereas the signal at

pH 7.0 might reflect the intermolecular tertiary con-
tacts. However, the information on three-dimensional
structure of EspA at different pH, particularly around
the amino acids between 60 and 147, which could not
be resolved by X-ray crystallography of EspA–CesA
complex, is critical to evaluate such a possibility.
Importantly, the pH 3.0, urea-induced unfolding
transition is almost superimposable onto the pH 7.0
transition curve. This observation suggests that the
protein, at pH 3.0, is as stable as that at pH 7.0. How-
ever, the ANS binding data indicate exposure of
hydrophobic surfaces at pH 3.0, probably due to
partial unfolding. One possible explanation, reconciling
this discrepancy, is that, unlike the traditional molten
globule, EspA maintains a well-ordered native-like
domain, but also has less structured regions with
exposed hydrophobic patches at pH 3.0. We designate
this conformational state, N
II
, the native structure at
acidic pH, which has a distinctive character compared
with the native conformation at neutral pH (N
I
). Thus,
the conformational change of EspA, associated with
changing pH, can be schematically represented as in
Scheme 1:
pH 2:03:05:0–7:0
I
A

… N
II
… N
I
The evidence suggests that the partially folded state at
pH 2.0 may have native-like tertiary contacts but a
lower a-helical structures content compared with N
I
and N
II
. It is now designated as I
A
, i.e. acid-induced
intermediate structure.
In an attempt to understand how pH and urea con-
centration affect the conformations of EspA, we con-
structed an EspA pseudo phase diagram with urea
concentration as a function of pH (Fig. 7), according
to Scheme 1. For ANS binding (Fig. 3), the ANS
transition midpoint can be considered to be the appar-
ent N
I
to N
II
transition midpoint, assuming that the
maximum ANS intensity in Fig. 3 corresponds to the
ANS fluorescence for N
I
. The urea-induced unfolding
transitions, between pH 3.0 and 5.0 (Fig. 6), provide

apparent midpoints for the transitions from either N
II
or N
I
to the unfolded state (U); whereas, the transition
Fig. 7. Pseudo phase diagram for EspA: urea concentration vs. pH
at 20 °C. The boundaries are defined by the ANS binding and the
urea-induced unfolding curves shown in Figs 3 and 5. U, Unfolded
state; I
A
, acid-induced intermediate state; N
I
, native state at neutral
pH; N
II
, native state at acidic pH. The transition midpoints for N
I
(or
N
II
) to U (circles), I
A
to U (squares) and N
I
to N
II
(triangles) are
shown by lines.
pH-Dependent EspA conformational change T. Kato et al.
2778 FEBS Journal 272 (2005) 2773–2783 ª 2005 FEBS

midpoint for I
A
to U is found using the pH 2.0 urea-
induced unfolding data. Importantly, since we have no
clear information on the transition between N
II
and N
I
by the addition of urea due to the spectral similarity
between these species, the boundary between N
I
and U
shown around neutral pH may actually correspond to
the boundary between N
II
and U. Also, unfortunately,
the experiments reported herein do not provide the
boundary between N
II
and I
A
. Additional experimenta-
tion using, for example, NMR or calorimetry is needed
to construct a more complete EspA phase diagram.
Although the phase diagram of Fig. 7 is incomplete, it
contains sufficient information such that, for a given
set of solution conditions, the existing conformational
state(s) can probably be identified.
The C-terminal regions (Val138 to Gln181) of two
EspA molecules may associate to form coiled-coil

structures [19]. These coiled-coils may then associate
further, forming oligomers. Based on our data, we pro-
pose that the oligomeric native state, found at neutral
pH, dissociates at pH 3 into a monomeric native-like
state with an ordered N-terminal domain and less
structured hydrophobic C-terminal tail.
The dissociation of oligomers into monomers upon
decreasing pH was previously observed for Salmonella
strain SJ25 flagellin [36]. In that case, the protein, at
acidic pH, assumes a conformation with an associated
ellipticity at 222 nm of )3800 deg °CÆm
-2
Ædmol
)1
.
Thus, some residual conformation may be present in
monomeric flagellin at acidic pH. It is possible that the
structural properties of monomeric flagellin, at acidic
pH, are similar to those of molten globules.
Oligomerization
The EspA filamentous superstructure has been ana-
lysed by electron microscopy [17,18]. It was suggested
that other factors, such as molecular chaperones, are
required to form an ordered EspA filamentous assem-
bly [18]. However, based on our sedimentation equili-
brium data, we suggest that recombinant EspA
spontaneously forms oligomers. For flagellin, several
additives, e.g. salts or polyethylenglycoles, are required
to induce formation of long, ordered filaments [37–41].
Unfortunately, we were unable to produce long EspA

filaments even when such additives were present (data
not shown). The results of the sedimentation equili-
brium experiment indicate that the largest oligomer
formed by the recombinant protein is approximately a
30-mer. According to the model derived from electron
microscopy, an axial rise for one filament is 4.6 A
˚
per
subunit [18]. Thus, a 30-mer, formed by recombinant
EspA, corresponds to a filament with a length of
approximately 14 nm. This is significantly shorter than
the length of EspA filaments formed on EHEC and
EPEC cell surfaces. Therefore, the assistance of addi-
tional factors, such as molecular chaperones, may be
needed to form longer EspA filaments, or the addi-
tional residues present at the N-terminal region of our
recombinant protein can destabilize the filaments.
Alternatively, time scales longer than those used in our
experiments may be necessary for the formation of suf-
ficiently long filaments.
In summary, we provide, herein, the first study con-
cerning the properties of the secondary structure of
EspA. EspA is shown to spontaneously associate into
oligomeric structures at neutral pH. However, two dis-
tinctive partially unfolded species occur at lower pH.
Based on these results, a phase diagram, illustrating
potential EspA conformational transitions, was con-
structed. Additional studies are necessary to character-
ize the EspA filamentous structure at the atomic level
and to elucidate the thermodynamic requirements for

filament formation. Such information should clarify
the role of EspA during host cell infection by EPEC
and EHEC.
Experimental procedures
Expression and purification of recombinant EspA
The espA gene was amplified from an E. coli O157:H7 cos-
mid library (RIMD 0509890, Sakai strain) [42,43] by PCR
and PCR product was cloned into pT7 vector. (Novagen,
Madison, WI, USA). The650 bp NdeI–SacI fragment con-
taining the espA gene was then inserted into the expression
vector, pET28a (Novagen). The recombinant EspA has
an additional 20 amino acids with the sequence
MGSSHHHHHHSSGLVPRGSH on the N-terminal side
of the native sequence. The plasmid pET28a-EspA was
transformed into E. coli BL21 (DE3).
Luria–Bertani (LB) broth, supplemented with 50 lgÆmL
)1
kanamycin, was inoculated with E. coli BL21 colonies and
incubated overnight at 37 °C with shaking. A portion of
the overnight culture was diluted 100-fold into fresh LB
medium and incubated at 37 ° C with shaking. Protein
expression was induced by addition of IPTG (at concentra-
tions up to 1 mm) when the cultures reached an optical
density of 0.5 at 600 nm.
After 3 h of further shaking at 37 °C, the cells were
harvested by centrifugation at 6000 g for 20 min at 4 °C
and the pellet was placed on ice for 15 min. Most
expressed EspA are located in insoluble fractions.
However, some EspA are also present in the soluble frac-
tion. Therefore, we prepared the recombinant EspA from

total cell solubilized by urea or from only soluble frac-
tions.
T. Kato et al. pH-Dependent EspA conformational change
FEBS Journal 272 (2005) 2773–2783 ª 2005 FEBS 2779
For preparation of EspA from urea-solubilized total
cells, the cells were resuspended in 100 mm sodium phos-
phate, pH 8.0, 10 mm Tris ⁄ HCl, 8.0 m urea and lysed by
sonication. The solution was centrifuged at 10 000 g for
30 min at 4 °C to separate the soluble and pellet fractions.
The soluble fraction was diluted drop-wise 100-fold into
50 mm sodium phosphate, pH 8.0, 300 mm NaCl, 10 mm
imidazole at 4 °C. The solution was loaded onto Ni–NTA
agarose (Qiagen, Valencia, CA, USA) and eluted using a
0–0.5-m imidazole gradient. The eluted EspA was dialysed
against 50 mm sodium phosphate pH 8.0, 300 mm NaCl,
10 mm imidazole and rechromatographed over Ni–NTA
agarose. Eluted EspA was concentrated by ultrafiltration
using a YM10 filter (Millipore, Billerica, MA, USA) and
then dialysed against 10 mm sodium phosphate pH 7.0.
Protein solutions were stored at )20 °C.
For purification from the soluble fraction, cells collected
by centrifugation were resuspended in 50 mm sodium
phosphate pH 8.0, 300 mm NaCl, 10 mm imidazole. Lyso-
zyme (1 mgÆmL
)1
final concentration) was added, and the
solution was incubated at 4 °C for 30 min. RibonucleaseA
and dideoxynuclease I (10 and 5 lgÆmL
)1
final concentra-

tions, respectively) were then added. Incubation was con-
tinued at 4 °C for a further 15 min. The solution was
cleared by centrifugation at 10 000 g at 4 °C for 30 min.
The supernatant was applied to Ni–NTA agarose equili-
brated with 50 mm sodium phosphate pH 8.0, 300 mm
NaCl, 20 mm imidazole, and washed with the same buffer.
The recombinant EspA was eluted with 50 mm sodium
phosphate pH 8.0, 300 mm NaCl, 250 mm imidazole. The
eluted protein was dialysed against, 50 mm sodium phos-
phate pH 8.0, 300 mm NaCl, 10 mm imidazole, and puri-
fied again by Ni–NTA agarose.
The purity of the recombinant protein was checked by
SDS ⁄ PAGE, which provided a single band around mole-
cular weight of 20 kDa, a value consistent with calculated
molecular weight of recombinant EspA. About 1 mg of
EspA were purified from 1 L culture by urea-solubilization
procedure, whereas only 0.1 mg of protein could purified
without solubilization by urea.
CD spectroscopy
CD spectra were recorded using a J-600 spectropolarimeter
(Jasco, Tokyo, Japan). The temperature was adjusted to
20 °C using a thermostatically controlled cell holder con-
nected to a circulating water bath. For far- and near-
UV CD spectra, cells of 1 mm and 1 cm path length were
used, respectively. Protein concentrations were 0.1 and
1mgÆmL
)1
for far- and near-UV CD measurements,
respectively. The samples were prepared about 12 h before
the measurements and the measurements were completed

within 24 h after preparation of samples. The sample pH
was checked by pH electrode, Horiba compact pH meter,
B-212 (Horiba, Kyoto, Japan) after each measurement. The
data were expressed as mean residue ellipticity, [h], where
[h] is defined as [h] ¼ 100 h
obs
(c · l)
)1
, h
obs
is the observed
intensity, c is the concentration in residue moles per litre,
and l is the path length in cm. The secondary structure
composition of EspA was estimated using the program
package CDPro [44,45]. Reported values are the average of
the results obtained from three independent programs: con-
tinll, selcon3 and cdsstr, according to the instruction of
cdpro program package. The [h] values between 200 and
250 nm with an interval of 0.2 nm taken at different pH
were directly used for input data.
The urea-induced unfolding curves were obtained by
plotting the ellipticity at 222 nm against urea concentration.
To estimate the urea concentration of midpoint of the
unfolding reaction (C
m
), the baseline for folded and unfol-
ded species are approximated from the plateau regions of
pre- and post-transition, respectively. The data were ana-
lysed according to the assumption of two-state transition
between a native and an unfolded state. However, we

should stress here that this analysis should be incorrect
because various oligomeric forms are present among native
conformers. However, without any data about the propor-
tion of each native oligmer, this treatment is the only the
probable and most conventional method to estimate C
m
values without any bias. The details in the analysis and the
parameters for unfolding are available as supplementary
material in Table S1.
Fluorescence spectrum
ANS fluorescence spectra were recorded using a FP-777
fluorimeter (Jasco). The excitation wavelength was 350 nm
and fluorescence emission spectra were recorded between
400 and 650 nm. The protein concentration was 0.1
mgÆmL
)1
(4.4 lm) and the ANS concentration was 5 lm.
The temperature was kept at 20 °C using a Peltier-type
thermostatically controlled cell holder.
Sedimentation equilibrium
Sedimentation equilibrium experiments were performed
using a Beckman Optima XL-I analytical ultracentrifuge
(Fullerton, CA, USA) operated at 15 000 r.p.m., 20 °C.
Various amounts of protein were dissolved in 20 mm
sodium phosphate pH 7.0, 100 mm NaCl. Using the pro-
gram, AA comp (RASMB web site: mb.
bbri.org/rasmb/mac/aa_comp-stafford), in conjunction with
the EspA amino acid composition, the partial specific vol-
ume of EspA was calculated as 0.731. The apparent
molecular weight (M

app
) was estimated according to the fol-
lowing equation:
M
app
¼
2RT
ð1 À tqÞx
2
d lnðCÞ
dðr
2
Þ
ð1Þ
pH-Dependent EspA conformational change T. Kato et al.
2780 FEBS Journal 272 (2005) 2773–2783 ª 2005 FEBS
where R is the gas constant, T is the absolute temperature,
x is the angular velocity, q is the solvent density and c is
the protein concentration at the radial distance r.
FT-IR
Infrared spectra were recorded using Avatar 370 (Thermo
Nicolet Co., Madison, WI, USA) under continuous purge
with dry nitrogen gas. Normal spectral resolution used
was 2 cm
)1
. The spectra of 128 scans were averaged. A
Happ–Genzel apodization function was applied before
Fourier transformation. The samples were transferred to
an IR sample cell consisting of a pair of CaF
2

windows
separated by a 15-lm spacer. FT-IR measurements were
carried out at room temperature. Recombinant protein
(5 mg) dissolved in 5 mL 10 mm sodium phosphate was
lyophilized and resuspended in 200 l L10mm sodium
phosphate ⁄
2
H
2
O at pH 7.0 or
2
H
2
O at pH 2.0. At pH 7.0,
visible precipitates were found in the solution. The spectra
of soluble and insoluble fractions were individually taken
after separating each fraction by centrifugation at
13 000 g at 4 °C for 20 min. The concentration of soluble
EspA at pH 7.0 was 1 mgÆmL
)1
estimated by UV absorp-
tion spectrum. At pH 2.0, no visible precipitates were
found. Thus, the concentration of EspA is considered to
be 10 mgÆmL
)1
. However, the presence of some aggregated
species was obvious from the FT-IR spectrum as discussed
in the text.
Acknowledgements
We thank Prof Yuji Goto for the use of the CD spec-

trometer and Miyo Sakai for performing the ultracen-
trifugation experiments. This work was supported in
part by grants-in-aid for scientific research from the
Japan Ministry of Education, Culture, Sports, Science
and Technology (MEXT).
References
1 Galan JE & Collmer A (1999) Type III secretion
machines: bacterial devices for protein delivery into host
cells. Science 284, 1322–1328.
2 McDaniel TK, Jarvis KG, Donnenberg MS & Kaper
JB (1995) A genetic locus of enterocyte effacement con-
served among diverse enterobactrial pathogens. Proc
Natl Acad Sci USA 92, 1664–1668.
3 McDaniel TK & Kaper JB (1997) A cloned pathogeni-
city island from enteropathogenic Escherichia coli con-
fers the attaching and effacing phenotype on K-12
E. coli. Mol Microbiol 23, 399–407.
4 Perna NT, Mayhew GF, Po
´
sfai G, Eliott SJ, Donnen-
berg MS, Kaper JB & Blattner FR (1998) Molecular
evolution of a pathogenicity island from enterohemor-
rhagic Escherichia coli O157: H7. Infect Immun 66,
3810–3817.
5 Zhu C, Agin TS, Elliott SJ, Johnson LA, Thate TE,
Kaper JB & Boedeker EC (2001) Complete nucleotide
sequence and analysis of the Locus of Enterocyte efface-
mant from rabbit diarrheagenic Escherichia coli RDEC-
1. Infect Immun 69, 2107–2115.
6 Shuch R & Maurelli AT (2000) The type III secretion

pathway. Dictating the outcome of bacterial–host inter-
actions. In Virulence Mechanisms of Bacterial Pathogens
(Brogden KA, Roth JA, Stanton TB, Bolin CA, Minion
FC & Wannemuehler MJ, eds), 3rd edn. ASM Press,
American Society for Microbiology, Washington, DC.
7 Clarke SC, Haigh RD, Freestone PP & Williams PH
(2003) Virulence of enteropathogenic Escherichia coli,a
global pathogen. Clin Microbiol Rev 16, 65–78.
8 Roe AJ, Hoey DE & Gally DL (2003) Regulation,
secretion and activity of type III-secreted proteins of
enterohaemorrhagic Escherichia coli O157. Biochem Soc
Trans 31, 98–103.
9 Kenny B, DeVinney R, Stein M, Reinscheid DJ, Frey
EA & Finlay BB (1997) Enteropathogenic E. coli
(EPEC) transfers its receptor for intimate adherence
into mammalian cells. Cell 91, 511–520.
10 Vlademir VC, Takahashi A, Yanagihara I, Akeda Y,
Imura K, Kodama T, Kono G, Sato Y & Honda T
(2001) Talin, a host cell protein, interacts directly with
the translocated intimin receptor, Tir, of enteropatho-
genic Escherichia coli, and is essential for pedestal for-
mation. Cell Microbiol 3, 745–751.
11 Vlademir VC, Takahashi A, Yanagihara I, Akeda Y,
Imura K, Kodama T, Kono G, Sato Y, Iida T & Honda
T (2002) Cortactin is necessary for F-actin accumulation
in pedestal structure induced by enteropathogenic
Eschericha coli infection. Infect Immun 70, 2206–2209.
12 Crane JK, McNamara BP & Donnenberg MS (2001)
Role of EspF in host cell death induced by
enteropathogenic Escherichia coli. Cell Microbiol 3,

197–211.
13 McNamara BP & Donnenberg MS (1998) A novel pro-
line-rich protein, EspF, is secreted from enteropatho-
genic Escherichia coli via the type III export pathway.
FEMS Microbiol Lett 166, 71–78.
14 Elliott SJ, Krejany EO, Mellies JL, Robins-Browne
RM, Sasakawa C & Kaper JB (2001) EspG, a novel
type III system-secreted protein from enteropathogenic
Escherichia coli with similarities to VirA of Shigella flex-
neri. Infect Immun 69, 4027–4033.
15 Kenny B & Jepson M (2000) Targeting of an entero-
pathogenic Escherichia coli (EPEC) effector protein to
host mitochondria. Cell Microbiol 2, 579–590.
16 Nougayre
`
de J-P, Fernandes PJ & Donnenberg MS
(2003) Adhesion of enteropathogenic Escherichia coli
to host cells. Cell Microbiol 5, 359–372.
T. Kato et al. pH-Dependent EspA conformational change
FEBS Journal 272 (2005) 2773–2783 ª 2005 FEBS 2781
17 Sekiya K, Ohishi M, Ogino T, Tamano K, Sasakawa
C & Abe A (2001) Supermolecular structure of the
enteropathogenic Escherichia coli type III secretion
system and its direct interaction with the EspA-sheath-
like structure. Proc Natl Acad Sci USA 98, 11638–
11643.
18 Daniell SJ, Kocsis E, Morris E, Knutton S, Booy FP &
Frankel G (2003) 3D structure of EspA filaments from
enteropathogenic Escherichia coli. Mol Microbiol 49,
301–308.

19 Delahay RM, Knutton S, Shaw RK, Hartland EL,
Pallen MJ & Frankel G (1999) The coiled-coil domain
of EspA is essential for the assembly of the type III
secretion translocon on the surface of enteropathogenic
Escherichia coli. J Biol Chem 274, 35969–35974.
20 Hyman HC & Trachtenberg S (1991) Point mutations
that lock Salmonella typhimurium flagellar filaments in
the straight right-handed and left-handed forms and
their relation to filament superhelicity. J Mol Biol 220,
79–88.
21 Knutton S, Rosenshine I, Pallen MJ, Nisan I, Neves
BC, Bain C, Wolff C, Dougan G & Frankel G (1998) A
novel EspA-associated surface organelle of enteropatho-
genic Escherichia coli involved in protein translocation
into epithelial cells. EMBO J 17, 2166–2176.
22 Ide T, Laarmann S, Greune L, Schillers H, Oberleithner
H & Schmidt MA (2001) Characterization of transloca-
tion pores inserted into plasma membranes by type III-
secreted Esp proteins of enteropathogenic Escherichia
coli. Cell Microbiol 3, 669–679.
23 Kodama T, Akeda Y, Kono G, Takahashi A, Imura K,
Iida T & Honda T (2002) The EspB protein of entero-
haemorrhagic Escherichia coli interacts directly with
a-catenin. Cell Microbiol 4, 213–222.
24 Chen Y-H, Yang JT & Martinez HM (1972) Determina-
tion of the secondary structures of proteins by circular
dichroism and optical rotatory dispersion. Biochemistry
11, 4120–4131.
25 Semisotnov GV, Rodionova NA, Kutyshenko VP,
Ebert B, Blanck J & Ptitsyn OB (1987) Sequential

mechanism of refolding of carbonic anhydrase B. FEBS
Lett 224, 9–13.
26 Ohgushi M & Wada A (1983) ‘Molten-globule state’: a
compact form of globular proteins with mobile side-
chains. FEBS Lett 164, 21–24.
27 Ohgushi M & Wada A (1984) Liquid-like state of side
chains at the intermediate stage of protein denaturation.
Adv Biophys 18, 75–90.
28 Ptitsyn OB (1995) How the molten globule became.
Trends Biochem Sci 20, 376–379.
29 Ptitsyn OB (1995) Molten globule and protein folding.
Adv Protein Chem 47, 83–229.
30 Arai M & Kuwajima K (2000) Role of the molten
globule state in protein folding. Adv Protein Chem 53,
209–282.
31 Creighton TE (1997) How important is the molten glob-
ule for correct protein folding? Trends Biochem Sci 22,
6–10.
32 Kuwajima K (1989) The molten globule state as a clue
for understanding the folding and cooperativity of glob-
ular-protein structure. Proteins 6, 87–103.
33 Kuwajima K (1992) Protein folding in vitro. Curr Opin
Biotechnol 3, 462–467.
34 Kuwajima K (1996) The molten globule state of a -lact-
albumin. FASEB J 10, 102–109.
35 Yip CK, Finlay BB & Strynadka NCJ (2005) Structural
characterization of a type III secretion system filament
protein in complex with its chaperone. Nature Struct
Mol Biol 12, 75–81.
36 Uratani Y, Asakura S & Imahori K (1972) A circular

dichroism study of Salmonella flagellin: evidence for
conformational change on polymerization. J Mol Biol
67, 85–98.
37 Novikov VV, Metlina AL & Poglazov BF (1994) A
study on the mechanism of polymerisation of Bacillus
brevis flagellin. Biochem Mol Biol Int 33, 723–728.
38 Abram D & Koffler M (1964) In vitro formation of
flagella-like filaments and other structures from flagellin.
J Mol Biol 116, 168–185.
39 Asakura S, Eguchi G & Iino T (1966) Salmonella
flagella: in vitro reconstruction and over-all shapes of
flagellar filaments. J Mol Biol 16, 302–316.
40 Asakura S, Eguchi G & Iino T (1964) Reconstitution of
bacterial flagella in vitro. J Mol Biol 10, 42–56.
41 Wakabayashi K, Hotani H & Asakura S (1969) Poly-
merization of Salmonella flagellin in the presence of high
concentrations of salts. Biochim Biophys Acta 175, 195–
203.
42 Hayashi T, Makino K, Ohnishi M, Kurokawa K, Ishii
K, Yokoyama K, Han CG, Ohtsubo E, Nakayama K,
Murata T, Tanaka M, Tobe T, Iida T, Takami H,
Honda T, Sasakawa C, Ogasawara N, Yasunaga T,
Kuhara S, Shiba T, Hattori M & Shinagawa H (2001)
Complete genome sequence of enterohemorrhagic
Escherichia coli O157: H7 and genomic comparison with
a laboratory strain K-12. DNA Res 8, 11–22.
43 Perna, NT, Plunkett G, 3rd Burland V, Mau B, Glas-
ner JD, Rose DJ, Mayhew GF, Evans PS, Gregor J,
Kirkpatrick HA, Posfai G, Hackett J, Klink S, Boutin
A, Shao Y, Miller L, Grotbeck EJ, Davis NW, Lim

A, Dimalanta ET, Potamousis KD, Apodaca J, Anan-
tharaman TS, Lin J, Yen G, Schwartz DC, Welch
RA & Blattner FR (2001) Genome sequence of
enterohaemorrhagic Escherichia coli O157: H7. Nature
409, 529–533.
44 Sreerama N & Woody RW (2000) Estimation of protein
secondary structure from circular dichroism spectra:
comparison of CONTIN, SELCON, and CDSSTR
methods with an expanded reference set. Anal Biochem
287, 252–260.
pH-Dependent EspA conformational change T. Kato et al.
2782 FEBS Journal 272 (2005) 2773–2783 ª 2005 FEBS
45 Sreerama N, Venyaminov SY & Woody RW (2000)
Estimation of protein secondary structure from circular
dichroism spectra: inclusion of denatured proteins with
native proteins in the analysis. Anal Biochem 287, 243–
251.
46 Rost B & Sander C (1993) Prediction of protein struc-
ture at better than 70% accuracy. J Mol Biol 232,
584–599.
47 Rost B & Sander C (1994) Combining evolutionary
information and neural networks to predict protein
secondary structure. Proteins 19, 55–72.
48 Rost B, Sander C & Schneider R (1994) PHD – an
Automatic Mail Server for Protein Secondary Structure
Prediction. CABIOS, 10, 53–60.
Supplementary material
The following material is available from http://www.
blackwellpublishing.com/products/journals/suppmat/EJB/
EJB4697/EJB4697sm.ht m

Table S1. C
m
values and apparent m
app
approximated
from urea-unfolding curves at different pH.
T. Kato et al. pH-Dependent EspA conformational change
FEBS Journal 272 (2005) 2773–2783 ª 2005 FEBS 2783