Embed Size (px)
Citation preview
http://www.elsevier.com/locate/bba
Biochimica et Biophysica
Effect of the polypeptide binding on the thermodynamic stability
of the substrate binding domain of the DnaK chaperone
Naoki Tanakaa,*, Shota Nakaoa, Jean Chatellierb, Yasushi Tania, Tomoko Tadaa, Shigeru Kunugia
aDepartment of Polymer Science and Engineering, Kyoto Institute of Technology, Matsugasaki, Sakyo, Kyoto 606-8585, JapanbShigaMedix, 192 Rue de la prairie, 63730 Les Martres de Veyre, France
Received 9 July 2004; received in revised form 8 November 2004; accepted 11 November 2004
Available online 19 January 2005
Abstract
The effect of polypeptide binding on the stability of the substrate binding domain of the molecular chaperone DnaK has been studied by
thermodynamic analysis. The calorimetric scan of the fragment of the substrate binding domain DnaK384–638, consisting of a h-domain and
an a-helical lid, showed two transitions centered at 56.2 and 76.0 8C. On the other hand, the thermal unfolding of the shorter fragment
DnaK386–561, which lacks half of the a-helical lid, exhibited a single transition at 57.0 8C. Therefore, the transition of DnaK384–638 at
56.2 8C is mainly attributed to the unfolding of the h-domain. The calorimetric scan of DnaK384–638D526N showed that the unfolding of
the h-domain was composed of two transitions. The polypeptide bound DnaK384–638 exhibited a symmetrical DSC peak at 58.6 8C,indicating that the substrate binding shifts the h-domain toward a single cooperative unit. A low concentration of GdnHCl (b1.0 M) induced
a conformational change in the h-domain of DnaK384–638 without changes in the secondary structure. While the thermal unfolding of the h-domain of DnaK384–638 was composed of two transitions in the presence of GdnHCl, the h-domain of the substrate bound DnaK384–638
exhibited a single symmetrical DSC peak in the same condition. All together, our results indicate that complex between DnaK384–638 and
substrate forms a rigid conformation in the h-domain.
D 2005 Published by Elsevier B.V.
Keywords: Molecular chaperone; DnaK; Substrate binding domain; DSC; Thermal unfolding; Limited proteolysis
1. Introduction
The Escherichia coli molecular chaperone DnaK, a
member of the HSP70 proteins family, facilitates the folding
of polypeptides and prevents their aggregation [1–3]. DnaK
consists of an N-terminal ATPase domain and a C-terminal
substrate binding domain (SBD). Substrate binding and
release is allosterically controlled by adenine nucleotides
[4]: When ATP is bound to the ATPase domain, the
substrate binding domain of DnaK (DnaK’s SBD) binds
substrate weakly (low affinity state), whereas ATP hydrol-
ysis brings about a conformational change in the ATPase
domain which in turn converts the SBD to a high affinity
1570-9639/$ - see front matter D 2005 Published by Elsevier B.V.
doi:10.1016/j.bbapap.2004.11.019
Abbreviations: SBD, substrate binding domain; GdnHCl, guanidine
hydrochloride; RCMLA, reduced and carboxylmethylated a-lactalbuminT Corresponding author. Tel.: +81 75 724 7861; fax: +81 75 724 7710.
E-mail address: [email protected] (N. Tanaka).
state. The 3D structures of several fragments of SBD have
been solved [5–10] as shown in Fig. 1. The structure of a
peptide–DnaK (residues 384–607) complex has been solved
by X-ray crystallography [5] and shows that the SBD
consists of a h-domain (393–501, consisting of strand
h1�8) and an a-helical lid (509–607, consisting of helices
A–E). The remaining C-terminal domain from residue 608
to 638 comprises random coil [10]. The deletion of this
region affects the ATPase binding, suggesting that this
region interacts to the ATPase domain to create a closed-to-
open equilibrium [11]. Substrate binding occurs by a
dynamic mechanism in a two-layered closing device
involving the independent action of an a-helical lid and
an arch formed by M404 and A429. The SBD is in
equilibrium between the open and closed conformations,
and the open conformation is largely populated in the ATP
bound states [4]. The NMR structure of SBD fragment
DnaK386–561 [6], in which helices C–E were deleted,
Acta 1748 (2005) 1–8
Fig. 1. 3D structure of the various DnaK’s SBD fragments. (a) DnaK389–
607 complexed with a model peptide (PDB code 1DKZ [5]). R445 and
D526, which form a salt-bridge [8], are shown in the ball-and-stick form.
The substrate model peptide is shown in green in the ball-and-stick form.
(b) DnaK386–561 (PDB code 1BPR [6]). R445 and D526 are shown in the
ball-and-stick form. (c) DnaK393–507 (PDB code 1DG4 [7]). All figures
were drawn with MOLSCRIPT [33] and Raster 3D [34].
N. Tanaka et al. / Biochimica et Biophysica Acta 1748 (2005) 1–82
suggests that the lid pivots on the h-domain during the
opening of the SBD (Fig. 1b).
The interaction between DnaK and its substrate has been
investigated in detail [12–19]. The NMR study of a lidless
SBD fragment DnaK393–507 revealed that the binding of a
substrate induces conformational changes in the h-domain in
absence of any interaction with the a-helical lid [7]. The
kinetic for the substrate binding of DnaK in the ADP bound
states is slow because of a large activation energy barrier [13].
DnaK384–638, a fragment of SBD, binds substrates with
high affinity in a similar manner to DnaK in the ADP bound
state [14,15,19], indicating that the majority of molecule is in
the closed conformation in the solution. The thermal
unfolding of the full-length DnaK [20,21], a fragment of
DnaK’s SBD (DnaK387–638) [21] and a fragment of human
HSP70’s SBD [22] has been described previously. These
studies revealed that the full-length DnaK and the fragments
of SBD exhibited multiple transitions in the thermal unfold-
ing. In this study, we have studied the effect of the
polypeptide binding on the thermal unfolding of DnaK’s
SBD fragments in order to gain insight into the effect of
substrate binding on its conformation. We found that the
substrate binding significantly changes the thermal unfolding
transition of DnaK’s SBD through the DSC measurement.
2. Materials and methods
2.1. Materials
The expression vector pDKC carrying DnaK384–638
with a 6xHis tag at the N-terminus was kindly provided by
Dr. W.F. Burkholder (MIT) and Prof. M.E. Gottesman
(Columbia University). The expression vector for
DnaK386–561 and DnaK386–507 was constructed by
PCR amplification and cloned into a pRSETA vector
(Invitrogen (Carlsbad, CA)). The cloned gene fragments
were sequenced to ensure that no mistakes had been
introduced during the amplification process. DnaK’s SBD
mutants were constructed using the quickchange site-
directed mutagenesis kit following the manufacturer instruc-
tions (Stratagene, La Jolla, CA). Recombinant proteins were
expressed as previously described [19], purified by affinity
chromatography using a NiNTA resin column following the
manufacturer’s instructions (Qiagen Inc., Valencia, CA).
The reduced and carboxylmethylated a-lactalbumin
(RCMLA) was obtained from Sigma Chemical Co. (St.
Louis, MO). 10 mM Tris pH 7.0 was used as the solvent for
the all experiments. Self-association of the fragment of the
substrate binding domain of DnaK has been reported in
previous studies [14,23] perhaps through the C-terminal
random coil region [11]. However, the oligomerization of
DnaK’s SBD was not monitored by native gel electro-
phoresis in our experimental condition (data not shown).
This would be due to the difference in the solvent condition
and the protein concentration.
2.2. DSC
Calorimetric measurements were performed on a Nano-
DSC II Model 6100 (Calorimetry Science Co., UT, USA).
Most experiments were done at a scan rate of 1.0 8C/min
and a protein concentration of 0.7–1.5 mg/ml. All data
analyses, i.e. baseline subtraction, concentration normal-
ization and deconvolution, were performed using the
software provided by the manufacturer (Calorimetry Sci-
ence Co., UT, USA).
The van’t Hoff enthalpy (DHvH) is obtained with the
standard formula,
DHvH ¼ 4 R T2mCp; max =DHcal
where Cp,max is the maximum of the excess heat capacity
function, Tm is the transition temperature defined as the
temperature location of Cp,max and R is the gas constant.
2.3. Limited proteolysis
Limited proteolysis was performed as described pre-
viously [24]. N-terminal sequence analysis was carried out
on the peptide samples isolated by blotting from a gel
using an Applied Biosystems (Foster City, CA) protein
sequencer (model 476A) equipped with an on-line analyzer
N. Tanaka et al. / Biochimica et Biophysica Acta 1748 (2005) 1–8 3
(model 610A) of phenylthiohydrantoin-derivatives of
amino acids.
2.4. Spectroscopy
The content of proteins secondary structure was moni-
tored by CD spectroscopic measurement using Jasco J-720
(Tokyo, Japan) at 1.0 AM protein concentration. Fluores-
cence measurements of ANS were preformed as described
previously [22] using a Shimadzu RF2000 spectrofluorom-
eter. The N-terminal amino group of RCMLA was labeled
with fluorescein isothiocyanate (FITC). The labeling ratio
for the fluorescent derivatives of RCMLAwas confirmed to
be equal to 1.0 from the concentration ratio of protein and
FITC. The molecular mass of FITC-RCMLAwas measured
using a Bruker Reflex III MALDI-TOFMAS in order to
confirm the equimolar reaction of FITC to RCMLA.
Fig. 3. Thermal transition of DnaK384–638 monitored by far-UV CD at a
protein concentration of 1.0 AM. The mean residue ellipticity at 222 nm of
DnaK384–638 in the temperature range from 25 to 90 8C was plotted. The
inset shows the intensity of the ANS-protein fluorescence with temperature
in arbitrary units (a.u.). Squares, ANS-DnaK384–638; dotted line, ANS-
human HSP70 [22].
3. Results
3.1. Thermal transition of the DnaK’s SBD fragments
We performed the calorimetric measurements of various
DnaK’s SBD fragments shown in Fig. 1. A calorimetric scan
of DnaK384–638 showed two thermal transitions at 56.2
and 76.0 8C (Fig. 2, line a). The transitions were reversible
after scanning to 90.0 8C. The thermal unfolding process of
DnaK384–638 is similar to those of the SBD fragments of
DnaK (DnaK387–638) [21] and human HSP70 [22]. The
calculated DHcal value for the transition at 56.2 8C of
DnaK384–638 is 434F12 kJ/mol, which is in the range of
Fig. 2. The temperature dependence of the partial heat capacity of the
various DnaK’s SBD fragments. (a) DnaK384–638; (b) DnaK386–561; (c)
DnaK384–638D526N; (d) DnaK386–507. For illustrative purposes the data
sets have been offset below the DnaK384–638 data set.
the previous value obtained for the SBD fragments of DnaK
and human HSP70. A calorimetric scan of DnaK386–561,
in which half of the a-helical lid is deleted, is shown in line
b of Fig. 2. In contrast to the thermal unfolding of
DnaK384–638, the transition became irreversible after
heating 90.0 8C, and no DSC peak was observed during
the consecutive scan. The irreversibility of the unfolding is
most likely due to protein aggregation, but the precipitate
was not observed. The calorimetric scan of DnaK386–561
showed only a single peak at 57.0 8C, indicating that the
peak at 56.2 8C of DnaK384–638 corresponds mainly to the
thermal transition of the h-domain of DnaK’s SBD. Since
the 76.0 8C transition of DnaK384–638 was not observed in
the calorimetric scan of the DnaK386–561 fragment, this
transition would be assigned to the thermal unfolding of the
a-helical lid.
To confirm this conclusion, we have monitored the
thermal unfolding of DnaK384–638 by CD spectroscopy.
Fig. 3 shows the thermal unfolding of DnaK384–638
monitored by the mean residue ellipticity at 222 nm. The
drastic change occurred from 60.0 to 80.0 8C, indicating thatthe unfolding of the a-helical lid mainly occurred in this
temperature region. Therefore, the transition of DnaK384–
638 at 76.0 8C monitored by DSC corresponds mainly to the
thermal unfolding of the a-helical lid. To characterize the
conformational state of DnaK384–638 at elevated temper-
ature, we have performed the fluorescence measurement of
ANS, which monitors the exposure of hydrophobic regions.
As shown in the squares plotted in the inset of Fig. 3, the
fluorescence of ANS was not increased in the process of
thermal unfolding of DnaK384–638 in the temperature
Fig. 4. The temperature dependence of the partial heat capacity of
DnaK384–638 in the presence of various concentrations of GdnHCl.
Dotted lines, DnaK384–638; solid lines, DnaK384–638+100 AM RCMLA.
Black line, in the absence of GdnHCl; red line, in the presence of 0.3 M
GdnHCl; blue line, in the presence of 0.5 M GdnHCl. We have confirmed
that more than ca. 70% of DnaK384–638 was saturated by RCMLA in this
experimental condition by the fluorescence titration method.
N. Tanaka et al. / Biochimica et Biophysica Acta 1748 (2005) 1–84
range from 25.0 to 75.0 8C. Therefore, the property of the
partly unfolded DnaK384–638 is not similar to the molten
globules.
The asymmetric shape of the peak at 56.2 8C for
DnaK384–638 indicates multiple transitions in the thermal
unfolding of the h-domain. The deconvolution analysis of
DnaK’s SBD in the previous study [21] showed that this
peak was composed of two transitions of roughly equal
DHcal centered at 50.4 8C and 58.2 8C. To verify this
finding, we have studied the thermal unfolding of the
mutant DnaK384–638D526N, in which the salt-bridge
between R445 and D526 [8] anchoring the a-helical lid
has been deleted. This salt-bridge stabilizes the closed
conformation of DnaK, and consequently, the D526N
mutation induces the open conformation mimicing the
effect of ATP in the full-length DnaK [25]. As shown in
line c of Fig. 2, the thermal unfolding of the h-domain of
DnaK384–638D526N is composed of two transitions. The
3D structures of DnaK’s SBD (Fig. 1) show that the h-domain consists of two h-sheets with four antiparallel h-strands: a first sheet of strands h3, h6, h7 and h8, and a
Table 1
Thermodynamic parameters associated with the thermal unfolding of DnaK384–6
Transition (1)
Tm (8C) DHcal (kJ/mol) DCp (k
DnaK384–638 56.2 434F12 8.7F0
DnaK384–638–RCMLA 58.6 544F10 13.2F0
DnaK386–561 57.0 665F5 6.4F0
second sheet of strands h1, h2, h4 and h5. The two peaks
observed in the calorimetric scan of DnaK384–638D526N
could be due to the different thermal unfolding transition
temperature of these two h-sheets.While the transitions of the a-helical lid and the h-
domain were independent in the thermal unfolding of
DnaK384–638, the truncated a-helical lid and the h-domain
unfolded simultaneously in the thermal unfolding of
DnaK386–561. This would be because conformations of
the h-domain are stabilized through the interaction with the
A and B helices. To confirm this result, we have examined
the thermal unfolding of the h-domain fragment DnaK386–
507, in which the a-helical lid was completely deleted.
Amazingly, the thermal transition peak centered at 57.0 8Cdid not appear in the temperature dependence of the partial
heat capacity of DnaK386–507, but the broad heat
absorption from 32.0 to 74.0 8C was observed (Fig. 2, line
d). The shape of the transition peak is too complicated to be
deconvoluted into some symmetry peaks. In addition,
precipitation accompanies the thermal unfolding of
DnaK386–507; precipitation is not observed for the thermal
unfolding of other DnaK’s SBD fragments. These results
indicate that the conformation of the h-domain alone tends
to aggregate and precipitate, and the a-helical lid plays a
role to prevent the self-association of the h-domain.
As shown in line b of Fig. 2, the shape of the thermal
transition of DnaK386–561 is symmetrical and DHvH/DHcal
value of the peak was calculated to be almost 1.0, indicating
that the thermal unfolding of DnaK386–561 is a two-state
transition [26]. Previous NMR study has revealed that the
substrate binding site of DnaK386–561 is occupied by its
own C-terminal tail [6], and this interaction would induce
the conformational change to form a single cooperative
folding unit. In order to test this hypothesis, calorimetric
experiments of DnaK384–638 were performed in the
presence and absence of excess RCMLA, the substrate
model unfolded protein (Fig. 4, black lines). The peak of the
h-domain of DnaK384–638–RCMLA complex is symmet-
ric and the calculated DHvH/DHcal value of the peak is
almost 1.0, indicating that the thermal unfolding is a two-
state transition (Fig. 4, solid line). The thermodynamic
characteristics of the thermal unfolding of DnaK’s SBD
fragments are summarized in Table 1. The DHcal value of
the thermal unfolding of the h-domain of DnaK384–638 at
56.2 8C is 434 kJ/mol. From the extrapolation using the DCp
value in Table 1, the DHcal value for the thermal unfolding
of the h-domain of DnaK384–638–RCMLA at 56.2 8C was
38 and DnaK386–561
Transition (2)
J/mol K) Tm (8C) DHcal (kJ/mol) DCp (kJ/mol K)
.4 75.8 312F10 6.8F2.2
.7 76.0 290F3 10.2F3.4
.3
N. Tanaka et al. / Biochimica et Biophysica Acta 1748 (2005) 1–8 5
calculated as 512 kJ/mol. The higher DHcal value for the h-domain of DnaK384–638–RCMLA demonstrates that the
binding of a substrate significantly increased the coopera-
tivity of the h-domain of DnaK384–638. Similarly, the
DHcal value of the thermal transition of DnaK386–561 at
56.2 8C was calculated as 660 kJ/mol, which is much higher
than the DHcal values for DnaK384–638–RCMLA at 56.2
8C. This would be due to the additional DHcal of the
unfolding of the truncated a-helical lid of DnaK386–561
accompanied with the thermal transition of the h-domain.
3.2. Thermal transition of the DnaK384–638 fragment in
the presence of low concentrations of GdnHCl
GdnHCl-induced denaturation of the full-length DnaK
has been investigated previously [27], and a stable
intermediate was observed from 0.6 to 1.0 M GdnHCl.
We have examined the effect of GdnHCl on the conforma-
tion of DnaK384–638 fragment by measuring the ellipticity
of far-UV CD spectra. An unfolding transition was observed
between 1.0 and 4.0 M GdnHCl, but no spectral changes
were observed at a concentration of GdnHCl lower than 1.0
M. This indicates that the secondary structures of
DnaK384–638 are not affected by the concentration of
GdnHCl lower than 1.0 M. We have then applied limited
proteolysis to the DnaK384–638 fragment in the presence of
0.5 M GdnHCl. An SDS-PAGE profile of the tryptic digest
of DnaK384–638 is shown in Fig. 5. The tryptic cleavage of
DnaK384–638 generated two fragments of ca. 28 kDa
shortly after initiating the digestion, and then two 16 kDa
fragments during a prolonged digest for 15 min. Most of the
band for non-digested DnaK384–638 disappeared after 30
min. The N-terminal amino acid sequences of the two ca. 28
kDa fragments correspond to residues 384–387 (GDVK)
and 388–391 (DVLL), indicating that the initial nicking
occurred in the labile N-terminal region. Both 16 kDa
fragments have the same N-terminus sequence, i.e. DAEA,
corresponding to residues 518–521, indicating that the
scissile peptide bond (R517–D518) is situated in a relatively
flexible region of DnaK384–638. The same region in full
length DnaK is highly susceptible to tryptic cleavage, and
Fig. 5. Time course of the SDS-PAGE profile of peptide fragments generated fro
terminated by acidification using 1% aqueous trifluoroacetic acid. SDS-PAGE w
buffer system.
was suggested to act as the hinge of the lid [28]. Conversely,
in the presence of excess RCMLA, DnaK384–638 was not
cleaved at the R517–D518 bond by trypsin (data not
shown). This is consistent with the conclusion from the
calorimetric experiment that the DnaK384–638–RCMLA
complex forms a rigid single cooperative conformation.
The SDS-PAGE profile of the trypsin digested
DnaK384–638 in the presence of 0.5 M GuHCl (Fig. 5)
shows additional fragment of ca. 20 kDa, with an N-
terminal sequence corresponding to residues 468–471
(GMPQ) of the h-domain. Therefore, the peptide bond
R467–G468 is made susceptible to proteolytic cleavage by
the presence of GdnHCl, indicating that the local con-
formation around R467–G468 was changed in the presence
of a pre-denaturation concentration of GdnHCl. In addition,
the proteolysis pattern obtained in the presence of GdnHCl
for the mutant DnaK384–638D526N is different (Fig. 5),
but the same R467–G468 site was also cleaved. A previous
study showed that the same tryptic cleavage was induced in
full-length DnaK by ATP binding [28]. These results
indicate that the local conformation of R467–G468 is
relatively flexible in the DnaK384–638 conformation and
is changed by a low concentration of GdnHCl.
The sensitive characterization of the energetics of the h-domain of DnaK384–638 was performed by recording
temperature scans in the presence of a low concentration
of GdnHCl [29]. The colored lines in Fig. 4 show the
calorimetric scan of DnaK384–638 in the presence of
GdnHCl. As shown in this figure, the Tm and DHcal values
for the transitions of the a-helical lid and the h-domain
decreased with the increase of the GdnHCl concentration.
This indicates that a low concentration of GdnHCl not only
changes the local conformation of the R467–G468 bond of
DnaK384–638, but also destabilizes the whole conforma-
tion of DnaK384–638. Fig. 6 shows the dependence of
DHcal on the Tm of the transition of the h-domain from the
different concentration of GdnHCl. When the unfolding is a
two-state transition, the slope of the function (dDHcal/dTm)
equals to the calorimetrically determined DCp [30]. The
open circle (o) of Fig. 6 shows that the DHcal vs. Tm plot for
DnaK384–638–RCMLA is a linear line, indicating that the
m the limited proteolysis of DnaK384–638 with trypsin. The reaction was
as performed using a slab gel with a concentration of 12.5% and a tricine
Fig. 6. Temperature dependence of the enthalpy changes associated with the
thermal unfolding of DnaK384–638. The Tm and the DH values of the
transition of the h-domain in the presence of various concentration of
GdnHCl were plotted. o, DnaK384–638+100 AM RCMLA in the presence
of 0 M, 0.3 M and 0.5 M GdnHCl; ., DnaK384–638 in the presence of 0
M, 0.1 M, 0.2 M, 0.3 M, 0.5 M and 1 M GdnHCl. The experimental
conditions are identical to those of Fig. 4.
N. Tanaka et al. / Biochimica et Biophysica Acta 1748 (2005) 1–86
h-domain of DnaK384–638–RCMLA complex undergoes a
cooperative two-state transition in the 0 to 1.0 M GdnHCl
concentration range. Consistently, the calculated slope for
DnaK384–638–RCMLA complex in the absence of
GdnHCl is 13.6F1.0 kJ/mol K, which is very close to the
DCp value of 13.2F0.7 kJ/mol K (Table 1). On the other
hand, the DHcal vs. Tm plot for the h-domain of DnaK384–
638 (.) was not linear, indicating that the thermal unfolding
of the h-domain of DnaK384–638 is not a two-state
transition. In addition, the shape of the transition was
drastically changed by 0.3 M GdnHCl as shown in red line
in Fig. 4. A previous study has shown that the full-length
DnaK is in a native conformation in the same GdnHCl
concentration range [27]. Therefore, the conformation of the
substrate binding domain of DnaK is more sensitive to the
effect from GdnHCl than DnaK molecules as a whole in the
absence of substrate polypeptide.
4. Discussion
We have characterized the thermal unfolding of the
various fragments of DnaK’s SBD. The calorimetric scan of
DnaK384–638 showed two thermal transitions at 56.2 and
76.0 8C. The thermal unfolding of the shorter fragment
DnaK386–561, which lacks half of the a-helical lid,
exhibits a single transition at 57.0 8C. Therefore, the peak
of DnaK384–638 at 56.2 8C is assigned as the thermal
unfolding of the h-domain. Since a drastic change in the
mean residue ellipticity at 222 nm of CD spectrum of
DnaK384–638 occurred from 60.0 to 80.0 8C, the DSC peak
at 76.0 8C is assigned as the thermal unfolding of the a-
helical lid. A previous study showed that the A–D helices of
DnaK form an antiparallel helical bundle, which is
stabilized by the interaction from the E-helix [11]. There-
fore, the A–E helices of DnaK form a cooperative folding
unit, which thermally unfolds at 76.0 8C.The thermal unfolding of DnaK384–638 at 56.2 8C was
composed of two transitions. This thermal unfolding process
is similar to that of HSP70’s SBD, in which two inter-
mediates are populated in the thermal unfolding transitions at
52.8 and 56.2 8C. The CD spectra of human HSP70’s SBD
showed that the secondary structure of these intermediates
was similar to that of the native conformation. However, the
intrinsic fluorescence of single tryptophan (W231) located
on the D-helix, which is buried in the native conformation,
showed that this residue was exposed to the solvent in the
intermediates. In addition, the fluorescence intensity of ANS
was increased when it bound to the intermediates of human
HSP70’s SBD in the thermal unfolding process (dotted line
in the inset of Fig. 3). These results indicate that the
spectroscopic property of the intermediates of human
HSP70’s SBD resembles to the molten globules. On the
other hand, the fluorescence of ANS was not increased in the
process of thermal unfolding of DnaK384–638, indicating
that the property of the partly unfolded conformation of
DnaK384–638 is not molten globule. This may be because
the A–E helices of DnaK384–638 form a cooperative folding
unit, and their conformations are not affected from the
unfolding of the h-domain. We found that the thermal
unfolding of DnaK384–638 was reversible, but the unfold-
ing of DnaK386–561 and DnaK386–507 was irreversible.
These results suggest that the a-helical lid plays a role to
prevent the aggregation of the h-domain.
The calorimetric scans of DnaK384–638 in the presence of
low concentration of GdnHCl showed that the conformation
of the substrate binding domain is more sensitive to the effect
from GdnHCl than DnaK molecules as a whole. Similar
results have been reported for the relationship between the
enzymatic activity and native conformation of proteins in the
presence of the low concentration of denaturant [31,32]:
When the inactivation and conformational changes are
measured in parallel in the enzyme, inactivation occurs at a
much lower concentration of denaturant than those requires to
trigger the unfolding of the enzyme molecule. As in the case
of the enzyme active site, the substrate binding domain of
DnaK would be in the rapid shifts in different conformations
to adopt the shape of the substrates.
We found that the polypeptide binding influenced the
energetics of the h-domain of DnaK384–638. The binding
of a polypeptide to the SBD shifts the two sheets toward a
single cooperative folding unit, and the DnaK384–638–
RCMLA complex undergoes a cooperative two-state ther-
mal unfolding even in the 0 to 1.0 M GdnHCl concentration
range. A previous DSC study of full-length DnaK reported
that the thermal unfolding of the h-domain at 58.0 8C of
N. Tanaka et al. / Biochimica et Biophysica Acta 1748 (2005) 1–8 7
SBD can be described by a two-state transition [21]. We
suggest that this would be also due to the self-binding of
polypeptide to the substrate binding domain: In the initial
stage of thermal unfolding of full-length DnaK, the ATPase
domain denatured first at 45.2 8C, and the unfolded ATPase
domain would bind to the substrate binding site intra-
molecularly to form a single cooperative folding unit in the
h-domain of SBD. Moreover, it has been shown that such a
self-binding has a huge effect on the conformation through
the ATPase activity of a lidless variant of DnaK [11]. ATP
reacts within the wild-type DnaK–peptide complexes
according to a two step reaction: ATP binds to the ATPase
domain in the first step, and the conformational change in
the ATPase domain triggering the close-to-open transition is
the substrate binding domain in the second step. ATP
induces the closed-to-open transition in the lidless DnaK
(DnaK1–517) with a first-order rate constant of 442 s�1,
whereas ATP triggers the closed-to-open transition in
DnaK1–554, which contains helices A and B, with a first-
order rate constant of 2.5 s�1. This large decrease has been
interpreted as the result of the self-binding of the B-helix to
the substrate binding site.
The effects of substrate binding on the conformation of
DnaK’s SBD have been extensively studied by NMR
[6,7,9]. The NMR solution structure of DnaK393–507, in
which the a-helical lid was completely deleted, is different
from the structure of the peptide bound DnaK substrate
binding domain. In the crystal structure of DnaK389–607
[5] and in the solution structure of DnaK386–561 [6], the
region Gln424–Ser427 of strand h3 is an internal part of the
lower h-sheet and is hydrogen bonded to strand h6.However, no characteristic cross strand NOEs was found
for these residues in the NMR structure of DnaK393–507,
indicating that this part of the sheet was not formed. The
extensive line broadening of their 1HN resonances was
observed in this region and for the amide protons of Gly405,
Thr417 and Ile418, which arises from the intermediate
exchange among multiple conformations. Such conforma-
tional flexibility was also observed in loop L2,3 which is
remote from the substrate binding cleft. Upon the addition
of the peptide to DnaK393–507, large chemical shift
changes occurred in the binding site region and residues
along strand h3 and loop L2,3. The final chemical shifts of
the peptide bound h-domain closely resemble those of the
intramolecularly bound 386–561. These results indicate that
the conformational exchange in the apo-h-domain is caused
by the lack of substrate, and substrate binding induced
structural and dynamic changes. These results are consistent
with our conclusion that the complex between DnaK384–
638 and substrate forms a rigid conformation.
Acknowledgments
We thank Prof. A.R. Fersht (Cambridge University) for
laboratory facilities, Dr. W.F. Burkholder (MIT) and Prof.
M.E. Gottesman (Columbia University) for pDKC and Mr.
T. Kuroita (Toyobo Co., Ltd) for technical assistance.
Financial support for this study was provided by the Asahi
Glass Foundation.
References
[1] B. Bukau, A.L. Horwich, The Hsp70 and Hsp60 chaperone machines,
Cell 92 (1998) 351–366.
[2] S.N. Witt, S.V. Slepenkov, Unraveling the kinetic mechanism of the
70-kDa molecular chaperones using fluorescence spectroscopic
methods, J. Fluoresc. 9 (1999) 281–293.
[3] F.U. Hartl, M. Hayer-Hartl, Molecular chaperones in the cytosol: from
nascent chain to folded protein, Science 295 (2002) 1852–1858.
[4] M.P. Mayer, H. Schrfder, S. Rudiger, K. Paal, T. Laugen, B. Bukau,Multistep mechanism of substrate binding determines chaperone
activity of Hsp70, Nat. Struct. Biol. 7 (2000) 586–592.
[5] X. Zhu, X. Zhao, W.F. Burkholder, A. Gragerov, C.M. Ogata, M.E.
Gottesman, W.A. Hendrickson, Structural analysis of substrate binding
by the molecular chaperone DnaK, Science 272 (1996) 1606–1614.
[6] H. Wang, A.V. Kurochkin, Y. Pang, W. Hu, G.C. Flynn, E.R.P.
Zuiderweg, NMR solution structure of the 21 kDa chaperone protein
DnaK substrate binding domain: a preview of chaperone–protein
interaction, Biochemistry 37 (1998) 7929–7940.
[7] M. Pellecchia, D.L. Montgomery, S.Y. Stevens, C.W. Vander Kooi,
H.P. Feng, L.M. Gierasch, E.R.P. Zuiderweg, Structural insights into
substrate binding by the molecular chaperone DnaK, Nat. Struct. Biol.
7 (2000) 298–303.
[8] R.C. Morshauser, W. Hu, H. Wang, Y. Pang, G.C. Flynn, E.R.P.
Zuiderweg, High-resolution solution structure of the 18 kDa substrate-
binding domain of the mammalian chaperone protein HSC70, J. Mol.
Biol. 289 (1999) 1387–1403.
[9] S.Y. Stevens, S. Cai, M. Pellecchia, E.P.P. Zuiderweg, The solution
structure of the bacterial HSP70 chaperone protein domain
DnaK(393–507) in complex with the peptide NRLLTG, Protein Sci.
12 (2003) 2588–2596.
[10] E.B. Bertelsen, H. Zhou, D.F. Lowry, G.C. Flynn, F.W. Dahlquist,
Topology and dynamics of the 10 kDa C-terminal domain of DnaK in
solution, Protein Sci. 8 (1999) 343–354.
[11] S.V. Slepenkov, B. Patchen, K.M. Peterson, S.N. Witt, Importance of
the D and E helices of the molecular chaperone DnaK for ATP binding
and substrate release, Biochemistry 42 (2003) 5867–5876.
[12] D.R. Palleros, L. Shi, K.L. Reid, A.L. Fink, Hsp70–protein
complexes. Complex stability and conformation of bound substrate
protein, J. Biol. Chem. 269 (1994) 13107–13114.
[13] C.D. Farr, F.J. Galiano, S.N. Witt, Large activation energy barriers to
chaperone–peptide complex formation and dissociation, Biochemistry
34 (1995) 15574–15582.
[14] W.F. Burkholder, X. Zhao, X. Zhu, W.A. Hendrickson, A. Gragerov,
M.E. Gottesman, Mutations in the C-terminal fragment of DnaK
affecting peptide binding, Proc. Natl. Acad. Sci. U. S. A. 93 (1996)
10632–10637.
[15] J. Zhang, G.C. Walker, Interactions of peptides with DnaK and C-
terminal DnaK fragments studied using fluorescent and radioactive
peptides, Arch. Biochem. Biophys. 356 (1998) 177–186.
[16] S.V. Slepenkov, S.N. Witt, Peptide-induced conformational changes
in the molecular chaperone DnaK, Biochemistry 37 (1998)
16749–16756.
[17] S.V. Slepenkov, S.N. Witt, Kinetic analysis of interdomain coupling
in a lidless variant of the molecular chaperone DnaK: DnaK’s lid
inhibits transition to the low affinity state, Biochemistry 41 (2002)
12224–12235.
[18] G. Buczynski, S.V. Slepenkov, M. Sehorn, S.N. Witt, Characterization
of a lidless form of the molecular chaperon DnaK, J. Biol. Chem. 276
(2001) 27231–27236.
N. Tanaka et al. / Biochimica et Biophysica Acta 1748 (2005) 1–88
[19] N. Tanaka, S. Nakao, D. Wadai, S. Ikeda, J. Chatellier, S. Kunugi, The
substrate binding domain of DnaK facilitates slow protein refolding,
Proc. Natl. Acad. Sci. U. S. A. 99 (2002) 15398–15403.
[20] D.R. Palleros, K.L. Reid, J. McCarty, G.C. Walker, A.L. Fink,
DnaK, hsp73, and their molten globules. Two different ways heat
shock proteins respond to heat, J. Biol. Chem. 267 (1992)
5279–5285.
[21] D. Montgomery, R. Jordan, R. McMacken, E. Freire, Thermodynamic
and structural analysis of the folding/unfolding transitions of the
Escherichia coli molecular chaperone DnaK, J. Mol. Biol. 232 (1993)
680–692.
[22] M.A. Fuertes, J.M. Perez, M. Soto, M. Menendez, C. Alonso,
Thermodynamic stability of the C-terminal domain of the human
inducible heat shock protein 70, Biochim. Biophys. Acta 1699 (2004)
45–56.
[23] M.P. Mayer, A. Buchberger, B. Bukau, Interaction of Hsp70
chaperones with substrates, Nat. Struct. Biol. 4 (1997) 342–349.
[24] N. Tanaka, C. Ikeda, K. Kanaori, K. Hiraga, T. Konno, S. Kunugi,
Pressure effect on the conformational fluctuation of apomyoglobin in
the native state, Biochemistry 39 (2000) 12063–12068.
[25] W.C. Suh, W.F. Burkholder, C.Z. Lu, X. Zhao, M.E. Gottesman,
C.A. Gross, Interaction of the Hsp70 molecular chaperone, DnaK,
with its cochaperone DnaJ, Proc. Natl. Acad. Sci. U. S. A. 95 (1998)
15223–15228.
[26] P.L. Privalov, Stability of proteins: small globular proteins, Adv.
Protein Chem. 33 (1979) 167–241.
[27] D.R. Palleros, L. Shi, K.L. Reid, A.L. Fink, Three-state denaturation
of DnaK induced by guanidine hydrochloride. Evidence for an
expandable intermediate, Biochemistry 32 (1993) 4314–4321.
[28] A. Buchberger, H. Theyssen, H. Schrfder, J.S. McCarty, G. Virgallita,
P. Milkereit, J. Reinstein, B. Bukau, Nucleotide-induced conforma-
tional changes in the ATPase and substrate binding domains of the
DnaK chaperone provide evidence for interdomain communication, J.
Biol. Chem. 270 (1995) 16903–16910.
[29] D. Xie, V. Bhakuni, E. Freire, Calorimetric determination of the
energetics of the molten globule intermediate in protein folding: apo-
a-lactalbumin, Biochemistry 30 (1991) 10673–10678.
[30] Y.V. Griko, P.L. Privalov, Calorimetric study of the heat and cold
denaturation of h-lactoglobulin, Biochemistry 31 (1992) 8810–8815.
[31] C.-L. Tsou, Location of the active sites of some enzyme in limited and
flexible molecular regions, TIBS 11 (1986) 427–429.
[32] C.-L. Tsou, Conformational flexibility of enzyme active sites, Science
262 (1993) 380–381.
[33] J. Kraulis, MOLSCRIPT: a program to produce both detailed and
schematic plots of protein structures, J. Appl. Crystallogr. 24 (1991)
946–950.
[34] E.A. Merritt, D.J. Bacon, Raster3D: photorealistic molecular graphics,
Methods Enzymol. 277 (1997) 505–524.
