Tài liệu Báo cáo khoa học: The conformational stability of the Streptomyces coelicolor histidine-phosphocarrier protein Characterization of cold denaturation and urea–protein interactions doc

Eur. J. Biochem. 271, 2165–2181 (2004) Ó FEBS 2004

doi:10.1111/j.1432-1033.2004.4142.x

The conformational stability of the Streptomyces coelicolor
histidine-phosphocarrier protein
Characterization of cold denaturation and urea–protein interactions
´
´
Jose L. Neira1,2 and Javier Gomez1
1

Instituto de Biologı´a Molecular y Celular, Universidad Miguel Herna´ndez, Elche (Alicante); 2Instituto de Biocomputacio´n y Fı´sica
de los Sistemas complejos, Zaragoza, Spain

Thermodynamic parameters describing the conformational
stability of the histidine-containing phosphocarrier protein
from Streptomyces coelicolor, scHPr, have been determined
by steady-state fluorescence measurements of isothermal
urea-denaturations, differential scanning calorimetry at
different guanidinium hydrochloride concentrations and,
independently, by far-UV circular dichroism measurements
of isothermal urea-denaturations, and thermal denaturations at fixed urea concentrations. The equilibrium unfolding transitions are described adequately by the two-state
model and they validate the linear free-energy extrapolation
model, over the large temperature range explored, and the
urea concentrations used. At moderate urea concentrations
(from 2 to 3 M), scHPr undergoes both high- and lowtemperature unfolding. The free-energy stability curves have
been obtained for the whole temperature range and values of
the thermodynamic parameters governing the heat- and
cold-denaturation processes have been obtained. Colddenaturation of the protein is the result of the combination
of an unusually high heat capacity change (1.4 ± 0.3

´
Correspondence to J. L. Neira and J. Gomez, Instituto de Biologı´ a
´
Molecular y Celular, Edificio Torregaitan, Universidad Miguel
´
Hernandez, Avda. del Ferrocarril s/n, 03202, Elche (Alicante), Spain.
Fax: + 34 966658459, + 34 966658459, Tel.: + 34 966658467,
E-mail: and
Abbreviations: CD, circular dichroism; DSC, differential scanning
calorimetry; Gdm Cl, guanidinium hydrochloride; DCp, the heat
capacity change; mDCpi , the heat capacity change upon preferential
urea-binding to the unfolded protein vs. the protein folded state; DHm,
the calorimetric enthalpy change at Tm; mDHi , the enthalpy change
upon preferential urea-binding to the unfolded protein vs. the protein
folded state; HPr, histidine phosphocarrier protein of the PTS;
scHPr, HPr from S. coelicolor; bsHPr, HPr from B. subtilis;
ecHPr, HPr from E. coli; LEM, linear extrapolation method;
PTS, the phosphoenolpyruvate-dependent sugar
phosphotransferase system; DSm, the calorimetric entropy
change at Tm; mDSi , the entropy change upon preferential
urea-binding to the unfolded protein vs. the protein folded
state; Tm, thermal denaturation midpoint.
´
Dedication: This paper is dedicated to the memory of Jose Laynez.
(Received 27 January 2004, revised 24 March 2004,
accepted 2 April 2004)

kcalỈmol)1ỈK)1, at 0 M urea, being the average of the fluorescence, circular dichroism and differential scanning calorimetry measurements) and a fairly low enthalpy change upon
unfolding at the midpoint temperature of heat-denaturation

(59 ± 4 kcalỈmol)1, the average of the fluorescence, circular
dichroism and differential scanning calorimetry measurements). The changes in enthalpy (mDHi ), entropy (mDSi ) and
heat capacity (mDCpi ), which occur upon preferential urea
binding to the unfolded state vs. the folded state of the
protein, have also been determined. The mDHi and the mDSi
are negative at low temperatures, but as the temperature is
increased, mDHi makes a less favourable contribution than
mDSi to the change in free energy upon urea binding. The
mDCpi is larger than those observed for other proteins; however, its contribution to the global heat capacity change upon
unfolding is small.
Keywords: calorimetry; denaturant–binding interactions;
histidine-phosphocarrier; protein stability.

A full understanding of the physical interactions underlying
the structure, folding and the function of a protein requires
a detailed description of its conformational stability in
terms of the free energy of unfolding. Such a thermodynamic description relies on the quantitative analysis of
denaturant-induced or thermally induced folding-unfolding
transitions, measured either spectroscopically or calorimetrically. In both cases, data analyses involves the extrapolation of the thermodynamic parameters to standard
conditions, usually 298 K in the absence of denaturant. To
extrapolate thermal denaturation data, the change in DCp,
and its temperature dependence must be known [1,2]. The
extrapolation of data from chemical-denaturation [with
either urea or guanidinium hydrochloride (Gdm Cl) as
denaturants] is carried out using either the linear free
energy model, LEM [3–5], or the binding model [6]. The
LEM is by far th e most commonly used model, and it has
been found to be valid for several proteins [7–9]. Combined
analysis of the LEM with thermal denaturation data,
assuming a temperature-independent DCp and the thermodynamic equivalence between the thermally and chemically

denatured states, have been reported for several proteins
[10, 7 and references therein]. These analyses yield the
thermodynamic parameters governing the conformational


´
2166 J. L. Neira and J. Gomez (Eur. J. Biochem. 271)

stability of the corresponding proteins, namely, the
enthalpy, entropy and the heat capacity changes. Recently,
other chemical-denaturation models have been proposed
based on: (a) a local-bulk partitioning, where the distribution of denaturant between the surface of the protein and
the bulk solution is described by a partition coefficient
[11,12] or (b) a model-independent nonlinear extrapolation
procedure which considers a truncated Taylor expansion of
the Gibbs energy function [13]. Both approaches have been
tested with model proteins and found to yield identical
conformational free-energies to those obtained by using the
LEM [11,13].
An essential step in the transport of carbohydrates across
the cell membrane of bacteria via the phosphoenolpyruvatedependent sugar phosphotransferase system (PTS) [14,15]
involves the transfer of a phosphoryl group from EI (enzyme
I of the PTS), the first protein in the cascade of proteins
forming the PTS, to HPr, the histidine-phosphocarrier
protein. HPr is the smallest protein in the protein cascade
of the PTS and it is thought to be a key element in the
regulation of PTS as it is always present in the phosphorylation of any sugar [16] and is involved in gene regulation [14].
The structures of HPr proteins from several species have been
described by NMR spectroscopy [17–19 and references
therein] and X-ray diffraction [20,21]. Those structures show

a classical open-face b-sandwich fold consisting of three
a-helices packed against a four-stranded antiparallel b-sheet.
This fold is also shown by proteins with no apparent
involvement in any phosphorylation reaction [22,23].
Streptomyces species are soil-dwelling actinomycetes
which grow on a variety of carbon sources, such as cellulose
and several types of mono- and di-saccharides. They are the
source of approximately two thirds of all natural antibiotics
currently produced by the pharmaceutical industry. The
complete genome of Streptomyces coelicolor has been
sequenced, showing the largest number of genes found in
any bacteria [24]. The presence of the different components
of the PTS in S. coelicolor has been reported, and the
corresponding proteins cloned and expressed [25–27]. HPr
of S. coelicolor, scHPr, contains 93 amino acid residues; it
lacks cysteine and tyrosine residues, and it only contains one
tryptophan and one phenylalanine residue. Assignment and
preliminary NMR studies of the HPr of S. coelicolor
indicate that its structure is similar to that observed in
other members of the HPr family (J. L. Neira, unpublished
results). As scHPr has a similar structure, but a different
amino acid sequence to HPrs of Escherichia coli, ecHPr,
or Bacillus subtilis, bsHPr, whose structures and folding
properties have been described previously [9,22,28–30], it is
important to understand whether the structure, sequence or
both determine the conformational stability in the HPr
family. There is a growing interest in determining to what
extent related proteins share the same conformational
stability features [31]. For instance, bsHPr seems to fold
via a two-state process [28], as was thought to occur also in

ecHPr [9,22]. Recently, however, the presence of non-native
contacts during ecHPr folding has been detected [30], and
structural rearrangements occurring upon folding around
an engineered tryptophan mutant have been observed [29].
Interestingly, scHPr seems to unfold at low pH via a
partially folded state [32]. The similarity between both
partially folded states in both HPr species remains to be

Ó FEBS 2004

elucidated, and it is also not yet known whether there is any
relationship between the presence of those states and the
different conformational stability between the HPr species
[32]. Thus, the study of the stability among the different HPr
members will allow one to establish whether there is a
common mechanism for the conformational stabilization in
this important family. In addition, the determination of
these conformational stabilities could provide the evaluation
of the thermodynamic parameters governing protein–
denaturant interactions, which, in turn, would shed light
on the still poorly understood mechanism of protein
denaturation. Attempting insight into those questions, we
use a two-part strategy in this work. First, we aimed to
determine the thermodynamic parameters governing the
conformational stability of scHPr (namely, DS, DH and
DCp), using different biophysical techniques [fluorescence,
circular dichroism (CD) and differential scanning calorimetry (DSC)] and to compare these with the thermodynamic
parameters obtained for other members of the HPr family,
that is, bsHPr (where only CD measurements were
performed) and ecHPr, where several biophysical techniques were also used by two independent groups [9,22]. The

use of different biophysical techniques allows comparison
between the different thermodynamic data obtained and,
thus, an assessment of the quality of the measurements.
Second, we aimed to determine the thermodynamic parameters governing the urea–scHPr interactions and their
temperature-dependence, and to compare them with those
obtained in other proteins.
Herein, it is shown that scHPr is only moderately stable
in aqueous solution. Its DG upon unfolding is only
4.0 kcalỈmol)1 at pH 7.5 at the temperature of maximum
stability. The analysis of the data performed at different
temperatures validate the LEM. The presence of moderate
concentrations of urea as a denaturant agent (2–3 M)
strongly destabilizes the native state of the protein with
cold-denaturation detectable at temperatures above 273 K.
The possibility to study both cold- and heat-denaturation
over a range of urea concentrations has made possible the
determination of the thermodynamic parameters governing
first, the HPr unfolding and, second, the urea–protein
interactions. The combination of denaturant and heatinduced denaturation experiments gave proof that cold
denaturation was a consequence of the combination of a
large heat capacity change (1.4 ± 0.3 kcalỈmol)1ỈK)1, at 0
M urea, being the average of the fluorescence, CD and DSC
measurements) and a low enthalpy change upon unfolding
at the midpoint temperature of heat-denaturation
(59 ± 4 kcalỈmol)1, the average of the fluorescence, CD
and DSC measurements). On the other hand, the enthalpy
and entropy changes upon preferential urea-binding to the
unfolded state vs. the folded state are negative at low
temperatures, but as the temperature is increased the
enthalpy makes a less favourable contribution than the

entropy to the free energy change upon urea–protein
interaction. Finally, the change in heat capacity and
enthalpy upon urea-binding is larger (116 ± 4 calỈmol)1Ỉ
K)1ỈM)1), than those observed in ecHPr [9] and bsHPr [28],
suggesting differential residual structure in the presence of
urea among the three proteins. However, the contribution
of mDCpi to the global heat capacity change upon unfolding
is small in the three proteins.


Ó FEBS 2004

Conformational stability of HPr of S. coelicolor (Eur. J. Biochem. 271) 2167

Experimental procedures
Materials
Ultra-pure urea used in fluorescence and CD, and the
Gdm Cl used in the DSC experiments were from ICN
Biochemicals. Imidazole, trizma acid, its base, and NaCl
were from Sigma. 2-Mercaptoethanol was from Bio-Rad,
and the Ni2+-resin was from Invitrogen. Dialysis tubing
was from Spectrapore (Los Angeles, CA, USA) with a
molecular mass cut-off of 3500 Da. Standard suppliers were
used for all other chemicals. Water was deionized and
purified using a Millipore system. Urea and Gdm Cl stock
solutions were prepared gravimetrically and filtered using
0.22 lm syringe-driven filters from Millipore. Exact concentrations of urea and Gdm Cl stock solutions were
calculated from the refractive index of the solution using an
Abbe 325 refractometer [33].
Protein expression and purification

The HPr clone comprises residues 1–93, with an extra
methionine and a His6-tag at the N terminus. We have
performed all the studies with this construction as its
structure, as observed by NMR (J. L. Neira, unpublished
results), is similar to that found in other members of the
HPr family and the His6-tag is disordered in solution,
making no contacts with the rest of the protein. Furthermore, stability measurements and biophysical characterization have shown no differences between the His-tagged
protein and that where the tag had been removed [32].
Recombinant protein was expressed and purified as
described elsewhere [32]. Protein was more than 99%
pure as judged by SDS protein-denaturing gels. Also,
mass spectrometry analysis was performed in a MALDITOF instrument, and only one peak was observed (data
not shown). The samples were dialysed extensively against
water and stored at )80 °C. Protein concentration was
calculated from the absorbance of stock solutions measured at 280 nm, using the extinction coefficients of model
compounds [34].
Fluorescence measurements
All fluorescence spectra were collected on a SLM 8000
spectrofluorometer (Spectronics Instruments, Urbana, IL,
USA), interfaced with a Haake water bath. A 0.5-cm pathlength quartz cell (Hellma) was used.
Urea-unfolding experiments were acquired by excitation
at 280 nm. The slit width was typically equal to 4 nm for
the excitation light, and 8 nm for the emission light. The
fluorescence experiments were recorded between 300 and
400 nm. The signal was acquired for 1 s and the wavelength
increment was set to 1 nm. Blank corrections were made in
all spectra. The unfolding curves were obtained in 10 mM
phosphate buffer, pH 7.5, either by direct titration of the
protein solution with urea or by preparation of different
solutions containing a constant concentration of protein

and different urea concentrations (between 0 and 6 M). Both
methods yielded superimposable sigmoidal plots for the
fraction of folded protein vs. urea concentration and
identical m- and transition midpoint-values.

Fluorescence spectra at different urea concentrations
were processed using the wavelength averaged emission
intensity, <k> [32].
Circular dichroism measurements
CD spectra were collected on a Jasco J810 spectropolarimeter fitted with a thermostated cell holder and interfaced
with a Neslab RTE-111 water bath. The instrument was
calibrated periodically with (+)10-camphorsulphonic acid.
Isothermal wavelength spectra at different urea concentrations (between 0 and 6 M) were acquired at a scan speed of
50 nmỈmin)1 with a response time of 2 s and averaged over
four scans at the desired temperature. Far-UV measurements were performed using 20–40 lM of protein in 10 mM
phosphate buffer (pH 7.5), using 0.1- or 0.2-cm path-length
cells. All spectra were corrected by subtracting the proper
baseline. The mean CD signal, [Q], was obtained from the
raw ellipticity data, Q [32].
Thermal-denaturation experiments were performed at
constant heating rates of 60 °CỈh)1 and a response time of
8 s. Measurements were acquired every 0.2 °C. Thermal
scans were collected in the far-UV region at 222 nm from
278 to 363 K in 0.1-cm path-length cells with a total protein
concentration of 20 lM. The reversibility of thermal transitions was tested by recording a new scan after cooling
down to 278 or 283 K the thermally denatured sample, and
comparing the thermal denaturation curve with that
obtained in the first scan. In all studies carried out here,
the experiments were fully reversible either for the heat- or
cold-denaturation processes. Thermal denaturation measurements were performed in the presence of different

amounts of urea ranging from 0 to 3 M (with maximum
temperatures of 363 K) and from 5 to 6 M (with maximum
temperatures of 323 K). Sample exposure to high temperatures was kept short to minimize any protein modification
by urea decomposition products and consequent irreversibility. The possibility of drifting of the CD spectropolarimeter was tested by running two samples containing
buffer, before and after the thermal experiments. No
difference was observed between both scans. In all cases,
after the reheating experiment, the samples were transparent
and no precipitation was observed. Care was taken to avoid
loss of volume due to evaporation by using a cap that sealed
the cuvette.
To asses the reproducibility of trends in the data and
fitted parameters, each of the CD measurements (either
thermal or chemical denaturation experiments) was repeated twice in two independent sets in the temperature range
explored. In all the experiments both set of data yielded
identical results.
Differential scanning calorimetry
DSC experiments were performed with a MicroCal MC-2
differential scanning calorimeter interfaced to a computer
equipped with a Data Translation DT-2801 A/D converter
board for instrument control and automatic data collection.
Lyophilized protein was dissolved in buffer (10 mM phosphate buffer, pH 7.5) and dialysed extensively against 2 L
of the same buffer at 277 K. All calorimetric experiments
were performed at concentrations of 1 mgỈmL)1. Samples


Ó FEBS 2004

´
2168 J. L. Neira and J. Gomez (Eur. J. Biochem. 271)

were degassed under vacuum for 10 min with gentle stirring
prior to being loaded onto the calorimetric cell. Samples
were heated at a constant scan rate of 60 °CỈh)1 and held
under a constant external pressure of 1 bar in order to avoid
both bubble formation and evaporation at high temperatures. Before rescanning, the samples were cooled in situ to
293 K for 40 min. Experimental data were corrected from
small mismatches between the two cells by subtracting a
buffer vs. buffer baseline, prior to data analysis. After
normalizing to concentration, a chemical baseline calculated
from the progress of the unfolding transition was subtracted. The excess heat capacity functions were then analysed
using the software package ORIGIN (Microcal Software, Inc.,
Northampton, MA, USA), supplied with the instrument.
For the experiments in the presence of Gdm Cl, a stock
solution of 6 M Gdm Cl concentration was used and
the corresponding amount of Gdm Cl was added. Gdm Cl
was used in the DSC measurements, instead of the urea
employed in the CD measurements (see above), to avoid
deamidation processes. Small concentrations of Gdm Cl
were used, because, as it has been shown [32], scHPr is
highly destabilized by the presence of Gdm Cl.
No differences were observed in the thermodynamic
parameters obtained when different scan rates were used
(data not shown and [32]).
Data analyses
Fitting of any equation described in this paragraph was
performed by using KALEIDAGRAPH (Synergy Software,
Reading, PA, USA) working on a PC.
Data analysis of the isothermal urea denaturation
curves. Far-UV CD and fluorescence chemical denaturation data were analysed using the two-state model for the
native (N) to denatured (U) protein equilibrium. According

to that model, the free energy governing the folding reaction
(DG) at a temperature T (in Kelvin), and the monitored
observable, X (either [Q] or <k>), are related [7–13, and
references therein] by:

(see below) and chemical-denaturations, the average emission intensity or the mean CD signal were converted to
plots of fU, the fraction of unfolded protein, which was then
given by
X XN T; ẵureaị
4ị
fU ẳ
XU T; ẵureaị XN T; ẵureaị
Thermodynamic equations either in the presence or in the
absence of chemical denaturant. For a two-state unfolding
reaction characterized by a temperature-independent heat
capacity change, DCp, within the temperature interval under
study, the equations for the dependence of the change in
enthalpy (DH), entropy (DS) and free energy (DG) are given
by [1,2,35]:
DHTị ẳ DHm ỵ DCp T Tm ị
5ị
 
T
6ị
DSTị ẳ DSm ỵ DCp ln
Tm



 

T
T
DGTị ẳ DHm 1
ỵ DCp T Tm T ln
Tm
Tm
7ị
In the above equations, Tm is the midpoint of the thermal
transition [i.e. the temperature at which DG(T) ¼ 0], which
is taken as the standard reference temperature. DHm and
DSm are the values of DH and DS at Tm, respectively.
Following the linear free-energy extrapolation model
[3–5,36] the changes in DH, DS, DG and DCp have a linear
dependence with denaturant concentration (the primes
denote the corresponding values of the thermodynamic
magnitude in the presence of urea):
DH0 ¼ DH mDHi ẵurea

8ị

DS0 ẳ DS mDSi ẵurea

9ị

DG0 ẳ DG mẵurea

10ị

DC0p ẳ DCp mDCpi ẵurea

11ị

DG=RTị

XN ỵ XU e
ị
Xẳ
1 þ eðÀDG=RTÞ Þ

ð1Þ

where XN and XU are the signals for the fully native and
fully unfolded states, respectively (the so-called baselines),
and correspond to the pre- and post-transition plateau
regions. The complete analysis of the thermal- (see below)
and urea-denaturation data requires an accurate determination of both baselines, which can be described as linear
functions of temperature (in K) and urea concentration
[7,28]:
XN T; ẵureaị ẳ X0N ỵ aN0 T ỵ bN0 ẵurea

2ị

XU T; ẵureaị ẳ X0U ỵ aU0 T ỵ bU0 ẵurea

3ị

where the rst term in each equation is the corresponding
observable value at 273 K in the absence of urea, for the
native and the unfolded states, respectively; the second term
is the linear slope of the observable change with the

temperature; and, the last term is the effect of urea on the
baselines. To allow for comparisons among the thermal-

where m, mDHi , mDSi and mDCpi are the changes in free
energy, enthalpy, entropy and heat capacity, respectively,
associated with the preferential interaction of urea with the
unfolded form of the protein relative to the folded form.
Assuming that mDCpi is temperature-independent, the temperature dependencies of m, mDHi , mDSi are then given by
[8]:
0
mDHi Tị ẳ mmi DHi ỵ mDCpi T Tm ị

 
T
mDSi Tị ẳ mmi DSi ỵ mDCpi ln
0
Tm


T
mi
mTị ¼ mDHi À TmDSi ¼ m DHi 1 À 0
Tm

 
T
0
ỵ mDCpi T Tm T ln
0
Tm

12ị
13ị

14ị

which are similar to Eqns 5, 6 and 7, respectively. Here,
mmiDHi and mmiDSi are the values of mDHi and mDSi at the


Ó FEBS 2004

Conformational stability of HPr of S. coelicolor (Eur. J. Biochem. 271) 2169

reference temperature Tm¢, which has been chosen as
the midpoint of the thermal denaturation (i.e. where
m(T) ¼ 0). Equation 14 indicates that the protein–urea
interactions are temperature-dependent.
The temperature dependencies of the free energy, DG¢,
the enthalpy, DH¢, and the entropy, DS¢, at any urea
concentrations are given by equations identical to Eqns
(7, 5 and 6), respectively. The temperature dependence of
DG¢ can also be described by the characteristic temperatures: Tm¢, Ts¢ and Th¢ [1,3], which are the temperatures
where DG¢, DS¢ and DH¢, respectively, are equal to zero.
The equations describing the relationships between those
characteristic temperatures, and DH¢ and DCp¢ are
described elsewhere [8,37].
Thermally induced denaturation curves monitored by farUV CD. The thermal denaturation curves obtained in the
presence of urea can be obtained by using Eqn (1) and
the expression of the DG¢ (which is analogous to Eqn 7).

The thermally induced denaturation data were converted to
plots of the fraction of protein in the unfolded state
according to Eqn (4). From this equation, the equilibrium
constant can be obtained in the folding transition region
and then DG¢ (i.e. the free energy at a given urea
concentration) is determined as a function of T (in K). A
plot of DG¢ vs. T at the melting temperature, TmÂ, yields a
slope equals to DHmÂ/Tm ẳ DSm¢, that is, the change in
entropy accompanying the unfolding reaction.
The temperature at which scHPr was denatured by
cooling is described in the literature [1,37].

Results
In scHPr, the spectroscopic and chromatographic studies,
and the coincident equilibrium unfolding curves obtained
with different spectroscopic probes [32] are consistent with a
two-state folding behaviour. The isothermal fluorescence
and far-UV CD urea-denaturation curves were, in all cases,
reversible. Isothermal urea denaturation curves were

Fig. 1. Urea-induced unfolding of scHPr monitored by the change in
intrinsic fluorescence spectra at 10 mM phosphate buffer, pH 7.5. In (A)
and (B), fU is plotted vs. the concentration of denaturant (urea) at
selected temperatures ranging from 278 to 323 K. The lines through
the data points represent the nonlinear least square fits to Eqn (1)
yielding the m- and [urea] half-values at each temperature. (C) Temperature dependence of the m-value from fluorescence measurements.
The errors bars are fitting errors to Eqn (1). The dotted line is the
linear temperature-dependence of the m-value, with a slope of
)8.6 ± 0.9 10)3 kcalỈmol)1ỈM)1ỈK)1. (D) Temperature dependence
of the Gibbs free-energy of unfolding. The solid line represents

the
nonlinear
least
square
fit
of
the  data
to:
h
i
T
DGðT Þ ¼ DH ðT0 Þ À T Á DSðT0 Þ þ DCp Á T À T0 À T Á ln T0 , which
is similar to Eqn (7) except that here T0, the reference temperature, was
taken as 298 K. At 298 K, the enthalpy, DH, entropy, DS, and free
energy changes, DG, upon unfolding of scHPr obtained from the fitting were 6.7 ± 0.5 kcalỈmol)1, 9.9 ± 1.5 cal mol)1ỈK)1 and
3.9 ± 0.2 kcalỈmol)1, respectively. The temperature dependence of
DG was consistent with a temperature-independent heat capacity
change, DCp, of 1.57 ± 0.29 kcalỈmol)1ỈK)1. The inset represents the
average energy obtained at 298 K.

obtained at 10 different constant temperatures from 278
to 323 K, when followed by fluorescence, and at eight
different temperatures from 278 to 318 K, when followed by
far-UV CD. However, because of the absence of a transition
when thermal-denaturation was followed by fluorescence
[32], the thermal-denaturation experiments were carried out
by observing the changes in ellipticity at 222 nm, using farUV CD.


Ó FEBS 2004

´
2170 J. L. Neira and J. Gomez (Eur. J. Biochem. 271)

Isothermal urea-denaturation monitored by changes
in the intrinsic fluorescence of the protein
Figure 1A,B shows fU as a function of urea concentration
for selected temperatures ranging from 278 to 323 K. The
unfolding mechanism was consistent with the two-state
model at all temperatures. Baselines were calculated from
Eqns (2 and 3) considering only the corresponding observ0
0
able value in the absence of urea (either XN or XU ), and the
urea coefficient (either bN0 or bU0 ) for the folded and
unfolded species, respectively. Each set of data was analysed
according to the LEM. The calculated free energy changes
upon protein unfolding are plotted vs. temperature in
Fig. 1D.
The m-values exhibited a slight tendency to decrease as
the temperature was raised from 1.43 ± 0.20 at 278 K to
0.98 ± 0.15 kcalỈmol)1ỈM)1 at 323 K. This decrease was
linear within the temperature range explored, yielding
a slope of )8.6 ± 0.9 · 10)3 kcalặmol)1ặM)1ặK)1 (Fig. 1C).
Conversely, the [urea]ẵ ([urea]ẵ is the urea concentration at
the transition midpoint) revealed a similar trend as
that observed for the temperature dependence of DG
(Fig. 1D).
As shown in Fig. 1D, the conformational stability of
scHPr was only moderate at neutral pH, reaching a
maximum of 4.0 ± 0.1 kcalỈmol)1. The free energy change

upon unfolding, DG, decreased both at higher and lower
temperatures. The temperature dependence of DG was
consistent with an enthalpy change, DH(298 K) of
6.7 ± 0.5 kcalỈmol)1, an entropy change, DS(298 K) of
9.9 ± 0.9 calỈmol)1ỈK)1 and a temperature-independent
heat capacity change equal to 1.57 ± 0.29 kcalỈmol)1ỈK)1,
which is in good agreement with that determined by the
analysis of the thermal- and urea-denaturation data
obtained by far UV-CD (1.05 ± 0.08 kcalỈmol)1ỈK)1 at
0 M urea, see below). The conformational stability vs.
temperature curve predicted a temperature of 259 K for the
cold-denaturation and 335 K for heat-denaturation. The
latter value is in good agreement with the results obtained

for the heat-induced denaturation experiments monitored
either by DSC (Tm ¼ 333.3 ± 3 K), by far-UV CD (Tm ¼
340 ± 2 K) (see below), and even the predicted colddenaturation temperature agrees with that determined by
far-UV CD (see below).
Finally, the temperature dependence of the conformational stability of scHPr by fluorescence reveals that the Th
was 294 ± 2 K, while the Ts was 296 ± 2 K (using the
equations described in the literature [8,37]). Both values are
in good agreement with those calculated from the heatdenaturation experiments in the presence of different urea
concentrations followed by far-UV CD (see below).
Heat-denaturation monitored by DSC
Figure 2 shows the excess heat capacity functions for the
heat-induced denaturation of scHPr in 10 mM phosphate
buffer, pH 7.5, in the presence of small quantities of Gdm
Cl, ranging from 0 to 0.2 M. The protein unfolds reversibly
via a two-state mechanism. The midpoint temperature of
the transition, Tm, as well as its enthalpy change upon

unfolding, DH(Tm), decreased as the concentration of Gdm
Cl increased from 0 [Tm ¼ 338 ± 3 K and DH(Tm) ¼
60 ± 2 kcalặmol)1] to 0.2 M [Tm ẳ 328 3 K and
DH(Tm) ẳ 47 3 kcalặmol)1). The unfolding of the
protein is consistent with a DCp value of 1.4 ± 0.2 kcalỈ
mol)1ỈK)1, which was determined from the slope of the
linear plot of DH(Tm) vs. Tm. This value is in good
agreement with that determined by far-UV CD at 0 M urea
(1.05 ± 0.08 kcalỈmol)1ỈK)1, at 0 M urea, see below) following the thermal- and chemical-denaturation curves and
with that determined from the fluorescence experiments,
following the urea-denaturation curves (see above).
The thermodynamic parameters for the unfolding of the
protein extrapolated at 298 K indicate that the conformational stability of the protein is only moderate (DG ¼
3.8 ± 0.3 kcalỈmol)1), the enthalpic contribution is still
favourable for the native state (DH ẳ 5.4 0.5 kcalặ
mol)1) and the entropic contribution unfavourable for the

Fig. 2. The excess heat capacity function of
scHPr at pH 7.5 in 10 mM phosphate buffer
containing small quantities of Gdm Cl as
destabilizing agent (0–0.2 M). In all the conditions tested, the protein was shown to unfold
reversibly by reheating the sample once it was
cooled down. The constant scanning rate was
60 °CỈh)1 and samples were heated up to
368 K. Both excess heat capacity functions,
from heating and re-heating scans, yielded
virtually identical Tm values, while the calorimetric enthalpy for the second scan was over
85% the value obtained for the first one. The
continuous lines represent the fitting of the
experimental data to a two-state reversible

model. Inset: temperature dependence of the
enthalpy change upon unfolding. In this temperature range, the unfolding of scHPr was
characterized by a temperature-independent
heat capacity change upon unfolding of
1.4 ± 0.2 kcalỈmol)1ỈK)1.


Ó FEBS 2004

Conformational stability of HPr of S. coelicolor (Eur. J. Biochem. 271) 2171

Fig. 3. Urea-induced unfolding of scHPr monitored by the change in
CD. (A) Urea-denaturation curves at selected temperatures at pH 7.5
(10 mM phosphate) as monitored by the change in ellipticity at 222 nm
in the far-UV CD spectra. The fraction of protein in the unfolded
form, fU, calculated using Eqn (4) is plotted as a function of urea
concentration at 278 K (s), 303 K (d) and 313 K (h). The inset
represents the changes in the raw ellipticity at 222 nm, 298 K. The
solid lines through the data are the nonlinear least squares fits to
Eqn (1). (B) Raw CD data at 298 K at different urea concentrations.
(C) Temperature dependence of the m-value from CD measurements.
The error bars are the fitting errors to Eqn (1). The dotted line is the
linear temperature-dependence of the m-value, with a slope of
)5 ± 3 · 10)3 kcalặmol)1ặM)1ặK)1. (D) The temperature dependence
of the [urea]ẵ (right side, s) and DG (left side, h). The error bars are
fitting errors to Eqn (1). The line through the DG data is the fitting to
Eqn (7). The errors are larger at the high temperatures, because the
pretransition regions were shorter.

folded state (–TDS ¼ )1.6 ± 0.3 kcalỈmol)1). These values

are in close agreement with those obtained from the
isothermal urea-denaturations followed either by fluorescence (see above) or far-UV CD (see below).
Isothermal urea-denaturations followed by far-UV CD
Experimental data at selected temperatures, plotted as the
fraction of unfolded protein, fU, are shown in Fig. 3A. Also
at 298 K, the raw data at selected urea concentrations are
shown (Fig. 3B). The m-values did show, over the examined
range, a slight linear temperature dependence, with a slope
of )5 ± 3 · 10)3 kcalỈmol)1ỈM)1ỈK)1 (Fig. 3C). The value
of this slope agrees, within the error, with that observed
in the urea-denaturations followed by fluorescence (see
above). The larger error in the CD measurements could be
due to the inherent larger errors (when compared to
fluorescence) obtained in the determination of the m-values
by using CD, as it has been shown in other proteins [38].
The [urea]½, which are more accurately determined and
therefore less susceptible to experimental errors than the
m-values, did show a temperature dependence (Fig. 3D)
similar to that observed for the free energy change of
unfolding in water, DG (Fig. 3D).
Thermal denaturation at fixed urea concentrations
as monitored by far-UV CD
Figure 4 illustrates the effects of urea on the thermostability
of the protein. At urea concentrations lower than 2 M,
scHPr showed a single conformational transition within the
temperature interval studied. Conversely, at urea concentrations larger or equal than 2 M, scHPr showed two
conformational transitions (both following a two-state
mechanism): one at low temperatures and other at high
temperatures, corresponding to cold- and heat-denaturation
of the protein, respectively. As urea concentration was

increased, the temperature for the heat-denaturation was
shifted at lower temperatures, while the midpoint temperature for the cold-denaturation increased (Fig. 4B). Above
3.5 M of urea, no significant thermal-transition was
observed (data not shown), which agrees with the results
obtained for isothermal chemical denaturation experiments

monitored by both steady-state fluorescence and far-UV
CD. These thermal denaturation data were analysed to
determine the DCp once the folded and unfolded baselines
were determined, as discussed below. The heat capacity
change is, in principle, assumed to be temperature-independent, although it changes to a small extent with
temperature [39,40] (see Discussion).
Determination of the pre- and post-transition regions.
The baseline for the fully folded protein in the CD
experiments was generated as follows. In 0 M urea, the
CD data in the pretransition region (278–310 K) were a


Ó FEBS 2004

´
2172 J. L. Neira and J. Gomez (Eur. J. Biochem. 271)

basically the same slope (the aU0 parameter in Eqn 3) with
a value of )0.025 ± 0.005 degỈcm)2Ỉdmol)1ỈK)1. The fits
among the data for the different urea concentrations
are then parallel to each other, with an offset corresponding to the urea contribution (the bU0 parameter in
Eqn 3). This value was added individually for each
thermal denaturation curve, including those carried out
at concentrations lower than 2 M urea where the posttransition baseline is not defined over a wide enough

temperature range (Fig. 4B). The coincidence of the slope of
the post-transition regions for the thermally and chemically
unfolded forms of the scHPr indicate that both unfolded
forms are thermodynamically equivalent. Similar findings
have been observed for both unfolded forms in other
proteins [7–9,28].

Fig. 4. Temperature- and urea-concentration-dependence of the mean
residue ellipticity at 222 nm. (A) s, 0 M); h, 1 M and d, 1.5 M. (B) s,
2 M; h, 2.5 M and d (3 M). Continuous lines through the data are the
fittings to Eqn (1), and the thermodynamic parameters governing such
transitions are given in Table 2.

function of temperature exclusively, and a linear fit provided
0
the intercept, XN , and temperature coefficient, aN0 , as
defined in Eqn (2). These two parameters were combined
with the pretransition CD data (278–310 K) obtained in
the presence of urea concentrations lower than 2.0 M to
evaluate the coefficient for the urea-dependent term. The
native baseline was then (Eqn 2):
HN T; ẵureaị ẳ 27:7 ặ 0:7ị ỵ 0:025 ặ 0:002ịT
ỵ 0:6 ặ 0:2ịẵurea
where HN T; ẵureaị is in units of degrecm)2Ỉdmol)1 at
222 nm, T is in K and [urea] is in M. The indicated errors in
the above expression are the fitting errors to Eqn (1). The
urea-dependent term, bN0 essentially shifted the baselines by
a constant amount in the thermal denaturation curves, and
it was very small for all the urea concentrations explored.
The above expression was used for all the thermal denaturation curves obtained, including those at 2, 2.5 and 3 M

urea, where the protein was either not completely folded at
any temperature (3 M) or was only folded for a small range
of temperatures (2 and 2.5 M) (Fig. 4B).
The baseline for the unfolded protein in the CD
experiments was obtained using the same approach described by other authors [7,8,36]. CD data for thermal
transitions in the presence of 2.5 or 3 M urea in the posttransition region (where the baseline was large enough,
Fig. 4B) and those obtained for the fully unfolded protein at
concentrations larger than 5 M urea, heated up to maximum
temperatures of 323 K (data not shown), were fitted
individually as linear functions of temperature, yielding

Determination of DCp. Once the native and unfolded
transition regions were determined for all the thermal
denaturation experiments, three different approaches were
used to determine DCp.
(a) Fitting of the CD thermal denaturation data at each
urea concentration to Eqn (1) yielded, for the heatdenaturation, the DHm¢ and Tm¢. These values were
estimated from a van’t Hoff analysis over a narrow
temperature range (usually lower than 5 °C), where the
unfolding transition occurs (i.e. for fU between 0.4 and 0.6).
All the thermal denaturation experiments were used in the
plot, except that of 3 M, where it was not possible to
determine the pretransition region as the protein was not
completely folded at any explored temperature (Fig. 4B).
The slope of a linear plot of DHm¢ vs. Tm¢ was the DCp
(a similar procedure has been used before in the determination of DCp from the DSC measurements). The linear fit
yielded a value of 1.3 ± 0.2 kcalỈmol)1ỈK)1 (Fig. 5A).
(b) For the 2, 2.5 and 3 M urea concentrations, the CD
thermal transitions revealed both heat- and cold-denaturations. In these cases, it is possible to obtain the complete free
energy stability curve as the curve of DG¢ changes its sign

twice (i.e. DG¢ equals zero twice). It can be shown that
Eqn (7) can also be written as [36]:
DC0p ỵ DS0 T0 ịị
lnK0ap ị ẳ
R
 
 

DC0p ỵ DS0 T0 ịị DG0 T0 ị
DC0p
T0
T0

ỵ

ln
RT0
R
T
R
T
15ị
where T0 is any chosen temperature reference. If this
temperature reference corresponded to either of the cold-,
0
T c m , or the heat-denaturation, Tm¢, temperatures (i.e. the
temperatures where DGÂ ẳ 0 kcalặmol)1), then the rst two
terms in Eqn (15) are equal but of opposite sign. Thus, if the
chosen temperature is Tm Eqn 15 is:
lnK0ap ị ẳ




DC0p þ DS0m Þ
ðÀDC0p þ DS0m Þ
Tm
þ
R
R
T
 
DC0p
Tm
ln
ð16Þ
À
R
T

The fitting parameters for the cold and heat-denaturation
data of 2.5 and 3.0 M urea are listed in Table 1, and Fig. 5B


Ó FEBS 2004

Conformational stability of HPr of S. coelicolor (Eur. J. Biochem. 271) 2173

more, they were similar to that determined using the other
two approaches (see before and the following paragraph).

(c) Pace and Laurents have described a method where it is
possible to obtain the DCp¢ by using the isothermal denaturation curves at any temperature, and the thermal denaturation data for any of the urea concentrations explored [39]. In
addition, the method also provides a validation of LEM.
Following that method, the entire DG¢ curve was determined
at any of the explored urea concentrations (from 0 to 2.5 M)
over a wide temperature range. The results from fitting the
experimental CD data with Eqn (7), at any urea concentration, with DHm¢, Tm¢ and DCp¢ as variable parameters, are
shown in Table 2 and Fig. 5C. Data at 3.0 M urea were not
taken into account because at this concentration, the folded
protein is not present at any explored temperature (Fig. 4B).
The DCp¢ at 0 M urea had the value of 1.05 ± 0.08 kcalỈ
mol)1ỈK)1, at 0 M urea, and the corresponding DCp (the
y-axis intercept in Eqn 11), is 990 ± 60 calỈmol)1ỈK)1. From
the data in Table 2, it seems that there was a slight trend in
DCp to increase with urea concentration, although this
tendency was small and fitting the data (DCp vs. urea
concentration) to Eqn (11) yielded a slope of 123 ± 40 calỈ
mol)1ỈK)1ỈM)1 (see Discussion). The value at 0 M urea
(1.05 ± 0.08 kcalỈmol)1ỈK)1) is in good agreement with that
determined from the van’t Hoff analysis and those determined by fluorescence and DSC (see before).
Determination of the Th¢ and Ts¢. The values of the
temperatures where the enthalpy and entropy are zero, Th¢
and Ts¢, respectively, were calculated by using equations
described in the literature [8,37] and are listed in Table 2. Th¢
was observed to increase with the urea concentrations
(Table 2). The vatiation of Th is given by:


dT 0 h
oDHm =DCp ị

ẳ
oẵurea
dẵurea
T
Fig. 5. Analysis of thermal unfolding curves of scHPr monitored by farUV CD signal at 222 nm at various urea concentrations, pH 7.5 (10 mM
phosphate). (A) DHm¢ vs. Tm¢ obtained by the vanÕt Hoff analysis of
thermal denaturation data. The error bars are fitting errors to Eqn (1).
(B) Analysis of thermal denaturation data at 2 M (s), 2.5 M (h) and
3 M (d) using the method of Chen and Schellman [36] as described in
the text. For the sake of clarity, only the data fit at 2.5 M to Eqn (16) is
shown (continuous line). (C) Fitting (solid lines) of DG¢ to Eqn (7)
according to the method of Pace and Laurents [39] for 0.5 M of urea
(j) and 1.5 M (d). Data at low temperatures were obtained by using
the LEM data (unfilled symbols), and those data at higher temperatures were obtained from the thermal denaturation experiments at the
specified urea concentration (filled symbols). Error bars are from the
fitting to Eqn (1).

shows the free-energy stability curves for those urea
concentrations and 2 M. The fittings for the unfolding at
both temperatures (cold and heat) at 2.5 and 3.0 M, yielded
the same DCp¢ (Table 1). Fitting of data at 2 M urea did not
yield good results for the cold-denaturation, probably
because this process was only observed at its early stages
(Fig. 5B). It is interesting to note that the DCP¢ obtained for
the heat-denaturation was the same, within the error,
among the three urea concentrations (Table 1). Further-

and then if DCp shows a small variation with urea
concentration, and thus, it can be assumed to be nearly
constant over the concentration range explored, the above

equation yields:


dT0h
1
oDHm
%À
DCp o½ureaŠ T
d½ureaŠ
In scHPr, the value of DCp is [denaturant]-dependent (see
Discussion), and the last approximation can not be strictly
applied; this shows why Th¢ increased, but also why it did
not change in a linear manner, even though there was a
linear relationship between DHm and [urea] (Fig. 5A).
Conversely, the Ts¢ remained constant, within the error,
at any urea concentration. This is due to the fact that Ts is
¢
predicted to increase nonlinearly with urea concentration,
according to:


dTs0
oðDSm =DCp ị
ẳ Ts0
oẵurea
dẵurea
T
In the region of 2.5 M urea, both temperatures become
equal; at this temperature (% 296 K) the fully folded and
unfolded states do not differ in enthalpy, entropy or in free

energy. Similar findings have been observed in ecHPr [9],
barstar [8], and a lac repressor DNA-binding domain [11].


Ó FEBS 2004

´
2174 J. L. Neira and J. Gomez (Eur. J. Biochem. 271)

Table 1. Thermodynamic parameters of the cold- and heat-denaturation of scHPr at different urea concentrations. Parameters were obtained by using
the method of Chen and Schellman [36] at pH 7.5 (10 mM phosphate). Errors are fitting errors to Eqn (1).
Heat-denaturation

Cold-denaturation

Urea (M)

Tm (K)

DCp (kcalặmol)1ặK)1)

DSm (calặmol)1ặK)1)

T c (K)
m

DC c (kcalặmol)1ặK)1)
p

DSm (calặmol)1ặK)1)

2a,b
2.5 b
3

322.9 ± 0.2
320.1 ± 0.2
306.0 ± 0.4

1.7 ± 0.3
1.43 ± 0.04
1.8 ± 0.1

142 ± 37
103 ± 46
33 ± 43

276.5 ± 0.5
295.4 ± 0.4

1.43 ± 0.04
1.8 ± 0.1

)108 ± 45
)30 ± 44

0

a

Attempts to determine the cold-denaturation at this urea concentration were unsuccessful, probably because of the absence of enough data
in the cold-denaturation region of the curve (Fig. 5B). b The values of the thermodynamic magnitudes for these two urea concentrations
agree, within the error, with those determined in Table 2.

Table 2. Thermodynamic parameters of the thermally induced denaturation of scHPr at pH 7.5 (10 mM phosphate) at fixed urea concentrations. Tm¢
was obtained from the fitting of the CD thermal denaturation data to Eqn (1). Fitting of the thermal and urea-denaturation data (using the
approach of Pace and Laurents [39]) yielded similar values, within the error. Tmc¢ was determined using equations described in the literature [1,37].
The errors are calculated from the propagation of fitting errors. DHm¢ , the enthalpy of the cold denaturation, was obtained from fitting to Eqn (1)
of the thermal denaturation CD data. Fitting of the thermal and urea-denaturation data (using the approach of Pace and Laurents [39]) yielded
similar values, within the error. Hm¢0 was determined from Eqn (5) with the value of DCp¢ and Tmc¢ listed in the table. DSm¢ and DSmc¢ were
0
c
m
calculated from the rates DHm or DHmc0 , respectively (see text). The errors are calculated from the propagation of fitting errors. Th¢ was determined by
T
Tm
using equations described in the literature [37]. The errors are calculated from the propagation of fitting errors. Ts¢ was determined by using
equations described in the literature [8,37]. The errors are calculated from the propagation of fitting errors. DCp¢ was obtained from fitting of the
thermal- and urea-denaturation data (using the approach of Pace and Laurents [39]). Indicated errors are fitting errors to Eqn (7) at different urea
concentrations.
Urea
(M)
0
0.5
1.0
1.5
2.0a
2.5a

340

333.3
332.6
327.6
322.2
313

Tm  (K)

Tm (K)







DHmÂ
(kcalặmol)1)

DHmcÂ
(kcalặmol)1)

DSmÂ
(calặmol-1ặK-1)

DSmcÂ
(calặmol)1ặK)1)

Th (K)

256
262
266
267
275
282

58
48
43
43
31
24

)58
)45
)40
)40
)29
)23

170
144
129
131
96
77

)228
)171

)153
)151
)106
)81

285
288
291
290
294
296

c

2
0.3
0.1
0.3
0.5
7








2
2

2
3
2
3

3
2
2
2
2
3








5
3
3
4
3
5

9
6
6
6
10
10








10
7
8
9
8
9

DCpÂ
(kcalặmol)1ặK)1)

Ts (K)
3
3
3
1
3
4

289
291
293
292
296
297

±
±
±
±
±
±

2
3
3
2

3
4

1.05
1.06
1.03
1.16
1.15
1.4

±
±
±
±
±
±

0.08
0.09
0.09
0.09
0.08
0.3

a

The values of the thermodynamic magnitudes for these two urea concentrations agree, within the error, with those determined by using the
approach of Chen and Schellman [36] ( Table 1).

Evaluation of the thermodynamical parameters

governing urea–protein interactions: m, mDHi and mDSi
By using Eqns (5–7), at different urea concentrations, and
the values of the DHm¢, DSm¢, Tm¢ and DCp¢ obtained by the
approach of Pace and Laurents [39] (Table 2), a detailed
analysis of the dependencies of DHm¢, DSm¢ and DCp¢ upon
urea concentration could be obtained. Only the thermodynamic parameters obtained from the analysis at 0, 0.5, 1
and 1.5 M urea concentrations were taken into account, due
to the errors associated in the determination of the free
energy curve at the highest urea concentrations (2.0, 2.5 and
3 M urea), where cold-denaturation was clearly observed
(Figs 4B and 5B).
In Fig. 6, the dependencies of DH¢ and DS¢ are shown
at two selected temperatures, 293 and 318 K. In both
cases, the errors in the determination of the thermodynamic parameters are large, due to the large scattering of
the measured data. At 293 K, DH¢ and DS¢ were
positive, and its absolute value increased linearly as urea
concentration changed. This suggests a favourable interaction of urea with scHPr at this temperature (Fig. 6A,B),

as it has been shown in other proteins [40]. The
compensation between both magnitudes leaded to a
resultant value of DG¢ that decreased linearly as the urea
concentration increased (Fig. 6C). Conversely, at 318 K,
DH¢ and DS¢ also had a positive value, but the absolute
magnitude decreased slightly as the concentration of urea
increased. Here, DG¢ also decreased linearly as urea
concentration increased, but it showed a better fit
(Fig. 6C) than those observed for DH¢ and DS¢. This
observation does not imply any thermodynamic feature
of the so-called enthalpy–entropy compensation as
besides the mainly artefactual nature of this correlation

[41], large errors in the determination of both DG and
DH have been invoked as the main reason why this
phenomenon is usually observed [42]. The behaviour of
DG¢ at the rest of the temperatures analysed was similar
to those described here for 293 and 318 K (data not
shown). It is worth mentioning here that: (a) the DG¢
values obtained from the linear fits agreed well with the
DG¢ values obtained directly from the LEM at the
chosen temperatures (data not shown) and (b) the slopes
of DG¢ agree, within the experimental uncertainty, with


Ó FEBS 2004

Conformational stability of HPr of S. coelicolor (Eur. J. Biochem. 271) 2175

Fig. 6. Urea concentration dependencies of DH¢ (A), DS ¢ (B) and DG¢
(C) at pH 7.5 (10 mM phosphate). The values for those three thermodynamic parameters are represented at 293 (h) and 318 K (s). The
straight lines through the data are linear least square fittings to Eqs (8–
10), respectively. To better appreciate the errors associated with the
experimental measurements, the plots of DH¢ (A) and DS¢ (B) show
the data on two different axis: 293 K on the left y-axis and 318 K on
the right y-axis. The y-axis intercepts and slopes of the lines are: (A) at
293 K, 5.1 ± 1.0 kcalỈmol)1 and 1.7 ± 1.0 kcalỈmol)1ỈM)1, respectively; and at 318 K, 33 ± 3 kcalỈmol)1 and 4.4 ± 3.9 kcalỈmol)1ỈM)1,
respectively; (B) at 293 K, 3.7 ± 3.3 calỈmol)1ỈK)1 and 0.7 ± 3 calỈ
mol)1ỈK)1ỈM)1, respectively; and, at 318 K, 98 ± 11 calỈmol)1ỈK)1
and 10 ± 12 calỈmol)1ỈK)1ỈM)1, respectively; (C), at 293 K,
3.5 ± 1.0 kcalỈmol)1, and 1.5 ± 0.8 kcalỈmol)1ỈM)1; and, at 318 K,
2.2 ± 1.8 kcalỈmol)1, and 1.2 0.9 kcalặmol)1ặM)1. The values of m
(the slope of DGÂ) agree, within the experimental uncertainty, with the

m-values determined from the isothermal denaturation experiments
(see text and Fig. 3C).

the slopes determined, from the fitting of fU to the urea
concentration (Eqn 1, Figs 1 and 3), that is the m-value
at the corresponding temperatures.
An analysis similar to that described in the previous
paragraph was performed at the eight chosen temperatures
in the range 278–318 K, and the values of mDHi mDSi and m
were determined (Fig. 7). The slope of the straight line of
mDHi vs. T yielded mDCpi which had a value of 115 ± 4 calỈ
mol)1ỈK)1ỈM)1. mDHi was small at very low temperatures

Fig. 7. Temperature-dependence of the thermodynamic parameters
which govern the interaction of scHPr with urea at pH 7.5 (10 mM
phosphate). (A) Temperature dependence of mDHi . The straight line
through the data is a linear least square fit of the data to Eqn (12),
whose slope is mDCpi and it has a value of 115.6 ± 0.5 calỈmol)1Ỉ
K)1ỈM)1. (B) Temperature dependence of mDSi . The straight line
through the data is a linear least square fit of the data to Eqn (13),
whose slope is mDCpi and it has a value of 115 ± 4 calỈmol)1ỈK)1ỈM)1.
(C) Temperature dependence of m. The solid line through the data is a
least square fit of the data to Eqn (14). The fitting yield a mDCpi equals
to 117 ± 2 calỈmol)1ỈK)1ỈM)1. The errors shown in the values of
mDHi , mDSi , and m are the error fits for the slopes in Eqns (8, 9 and 10),
respectively.

(278 K), reaching the zero at 279 K, and becoming
increasingly positive as the temperature was raised
(Fig. 7B). mDSi was also small at low temperatures, and it

made favourable contributions to m above 289 K (Fig. 7C).
Both sets of thermodynamic magnitudes resulted in a free
energy curve which had a maximum, as it was expected
from Eqn (14). The fitting to this curve yielded a value of
mDCpi of 117 ± 2 calỈmol)1ỈK)1ỈM)1, which agreed, within
the error, with that determined from the variation of mDHi
with the temperature. Also, both values agree, although the
experimental uncertainty is rather large, with the value


´
2176 J. L. Neira and J. Gomez (Eur. J. Biochem. 271)

determined from the variation of DCp¢ vs. urea concentration (Eqn 11), which was 123 ± 40 calỈmol)1ỈK)1ỈM)1 (see
above).

Discussion
The conformational stability and the cold-denaturation
of scHPr
The conformational stability of a protein, DG, is fully
specified when the enthalpy and the entropy changes at a
chosen reference temperature (DHm and DSm, respectively),
and the heat capacity changes (DCp) are known. By using a
combination of isothermal urea- and thermal-denaturation
experiments followed by CD, fluorescence and DSC, it has
been possible to obtain the stability curve of scHPr at
different urea concentrations and over a wide temperature
range. It is important to note here that three different
biophysical techniques were used in this work to assess the
reproducibility of trends in the experimental data and the

fitted parameters.
Several proteins have been shown to undergo cold
denaturation, such as the monomeric k repressor [43],
CheY [44], the human fibroblast growth factor [45], other
HPr family members [9,22,28], 434 Cro protein [7] and
barstar ([8], and references therein). The main difficulty to
detect the cold-denaturation has been the temperature at
which it happens. According to the thermodynamic equations [1,37], cold denaturation should occur at a temperature midpoint, T m c0 , below the freezing point of water for
most proteins. Then, to observe cold denaturation above
the freezing point of water, proteins need their native state
to be previously destabilized. This means that the study of
cold-denaturation is greatly facilitated by the presence of
denaturants (chemicals or pH), as denaturants raise the
cold-transition temperature and depress the freezing point
of water [1,8,9,22,28]. In the study of the stability of scHPr
shown here, urea has been used to destabilize the folded
state.
The temperature of maximal stability in scHPr, Ts, was
292 ± 4 K (Table 2) (the average of the values obtained at
the different urea concentrations). The stability of scHPr
decreased as the temperature was lowered (or increased)
below (above) 292 K, at any urea concentration (Fig. 5C).
The decrease in stability at low temperatures observed in the
presence of urea can be rationalized by the changes in the
thermodynamic parameters governing the urea–protein
interactions. Briefly, the free energy at any urea concentration, DG¢, depends on the conformational free energy of the
isolated protein, DG, and the free energy corresponding to
the urea–protein interactions, m, according to Eqn (10). As,
at low temperatures, mDHi is small (Fig. 7A), making
favourable contributions to m (that is, the urea–protein

interactions are favoured, Eqn 14), it can be concluded that
at low temperatures, DG¢ is high at any urea concentration.
Comparison among the stability curves at different urea
concentrations indicates that the free energy was also
decreased when urea concentration was raised (Fig. 5C),
due to the increase in the number of urea–protein interactions.
For concentrations < 2.0 M destabilization of the native
state of the protein was not enough to detect a significant

Ó FEBS 2004

structural unfolding transition at low temperatures
(Fig. 4A) and only a transition at high temperatures
(heat-induced) was observed. However, at urea concentrations ‡ 2.0 M two transitions were observed: one at high
temperatures (heat-induced) and other at lower temperatures (cold-induced) (Fig. 4B). Some conclusions can be
drawn from the calculation of the thermodynamic parameters governing both transitions (Table 2). First, it can be
observed that as the urea concentration was increased the
Tm¢ was reduced and, concomitantly, the T m c0 was increased.
Also, it is clear from Table 2 that cold-denaturation is
accompanied by a substantial decrease in entropy. As it is
also clear that at 0 M urea the folded scHPr cannot have a
higher conformational entropy than the cold-denatured
protein, the large decrease in the entropy upon colddenaturation must be accounted for by a change in the
entropy of the water. This suggests that the hydrophobic
effect, as it happens in barstar [8], ecHPr [9,22] and bsHPr
[28] is the dominant force stabilizing scHPr. On the other
hand, it can also be seen from Table 2 that DH m c0 had a
negative value, which was similar, but of opposite sign to
that observed in the heat-denaturation. As it has been
suggested [8], this decrease in the enthalpy upon colddenaturation must be due to the change in the enthalpy of

interaction of water.
Why is cold-denaturation observed in scHPr? It has been
concluded [1,2,7–9,22,28] that the observation of colddenaturation relies on the magnitude of the DCp and
DH(298 K), which are the thermodynamic parameters
involved in the determination of the cold-denaturation
temperature [1,37]. Proteins undergoing cold-denaturation
have smaller DH(298 K) and larger DCp values than a
protein of similar size. The DH(298 K) of scHPr is
10 kcalỈmol)1; this value is smaller than those found in
other proteins, which range from 20 to 80 kcalỈmol)1 (see
[7–9,28] and references therein). Furthermore, the DCp of
scHPr is 1.4 ± 0.3 kcalỈmol)1ỈK)1 (the average of the DSC,
CD at 0 M urea and fluorescence measurements), which
yields a value per residue (taken into account that the Histag is disordered in solution, as shown by NMR (J. L. Neira,
unpublished results), and that it does not contribute to the
stability of the protein [32]), of 14 calỈmol)1ỈK)1, which is as
high as that measured in the 434 Cro protein [7], where colddenaturation has also been observed. However, that value is
not as high as those observed in barstar [8] or other HPrs.
Cold denaturation was also observed in bsHPr [28] and
ecHPr [9,22] and these processes were explained in similar
terms to those described here (a small DH and a large DCp).
In those HPrs, the complete set of thermodynamic parameters were determined using only the CD measurements, for
bsHPr [29], and by using CD [9,22], fluorescence [22] and
DSC measurements [22] for ecHPr. Among the three HPr
members, the DHm are similar for scHPr and bsHPr
(59 ± 4 in scHPr, the average of the DSC, the CD and
fluorescence measurements, vs. 59 ± 2 kcalỈmol)1 in bsHPr
[28]), but very different to that observed in ecHPr
(76 ± 3 kcalỈmol)1 [9], and 69 ± 1 kcalỈmol)1 [22]). Conversely, the DCp values are the same, within the error,
between ecHPr and scHPr (1.4 ± 0.3 in scHPr vs.

1.45 ± 0.08 in ecHPr [9]); but they are different to that
observed in bsHPr (1.16 ± 0.05 kcalỈmol)1ỈK)1 [28]),
although it is within the error of the value measured in


Ó FEBS 2004

Conformational stability of HPr of S. coelicolor (Eur. J. Biochem. 271) 2177

scHPr. The larger error in the DCp of scHPr is due to the
fact that it is the average of the three different measurements
(DSC, CD and fluorescence measurements).
As a consequence of the large DCp and small DH the
stability curve of scHPr at 0 M urea also shows an unusual
feature. The curve has a relatively high Tm (Table 2) and a
small conformational stability at 298 K (% 4 kcalỈmol)1,
Figs 1C and 3C), at Ts. Thus, a large DCp should sharpen
the stability curve (and then a low Tm should be expected)
[46–48]. The small conformational stability at Ts can be
explained by the large measured value of DCp, as:
DG(Ts) ¼ DHm – (Tm ) Ts) DCp [37,49]. To explain the
larger Tm measured, the equation which allows the determination of Ts must be used [8,37]:


DHm
Ts ¼ Tm exp À
Tm DCp
If Ts must be around room temperature, as it has been
concluded from a statistical survey of the thermodynamic
parameters of several proteins [47] and as it has been

found here for scHPr (Table 2), the small DHm, the large
DCp and the exponential function make Tm shift towards
higher values (but not very high as Tm is also in the
denominator of the exponential). This displacement of the
Tm to higher values has also been explained by considering
the changes in water structure upon heating and [50,51].
The determination of DCp and its variation with urea
and temperature
Evaluation of the DCp from stability data requires the
determination of the second derivative of the stability from
the experimental measurements. The possible conclusions
should be interpreted with these experimental limitations in
mind. We describe, first, the temperature and urea dependence of DCp, and, second, this parameter is compared with
those found in other proteins.
We cannot rule out the temperature-dependence of the
heat capacity based on our data. The use of the approach
developed by Pace and Laurents [39] yields a precise
estimation of DCp¢ (Table 2 and Fig. 5C) as the slope of
the free energy curve changes from a large negative value
to a large positive value upon decreasing temperature.
However, this does not prove the temperature-independence of the heat capacity as that dependence is provided
by the free energy equation (Eqn 7). The use of the Chen
and Schellman approach [36] to determine the DCp¢ at the
cold- and heat-denaturation temperatures (Table 1 and
Fig. 5B) and the fact that the same DCp¢ value was
observed for both denaturations (Table 1) does not prove
either the temperature-independence of DCp¢, as Eqns (15
and 16) are obtained on such an assumption [36]. The
consistency of the whole thermodynamic data seem to
support, within the error of the experimental measurements, such assumption, but we cannot rule out a small

temperature-dependence in the heat capacity, as it has
been observed in other proteins [37,39,52]. In fact, it has
been suggested that the temperature-dependence of DCp
follows a broad bell-shaped function with a maximal
constant value around 313 K [53], based on the curvature
in the heat capacity curve found at lower temperatures for
the unfolded state [54].

On the other hand, it is possible to conclude the ureadependence of DCp within the experimental uncertainties,
based first, on the experimental variation of DCp with urea
(Table 2); and, second, on the agreement between the slope
obtained from that variation and the mDCpi determined from
the variations of mDHi , mDSi and m with temperature
(Fig. 7). However, those slopes are small, 116 ± 4 calỈ
mol)1ỈM)1ỈK)1 (the average obtained from the mDHi and m
curves) when compared to the DCp. If DCp increases with
urea concentration (consistent with a positive mDCpi ) a heat
capacity increase should be expected to follow the transfer
of model compounds from water to urea solutions. Several
studies have tried to address this question [55–59]. These
studies suggest that: (a) the transferral from water to urea of
nonpolar groups results in a decrease in the heat capacity,
being the process enthalpically disfavoured; (b) transferral
of polar groups (from water to urea) results in an increase in
the heat capacity, and the process is enthalpically favoured;
and (c) the transferral (from water to urea) of the
polypeptide backbone has either an increase or nearly zero
value in the heat capacity. T hese results can be rationalized
considering that, upon transferral to urea, there is a
disruption of solvent structure around nonpolar groups

and formation of solvent structure around the polar groups
and the polypeptide backbone. It is possible that in scHPr,
upon addition of urea, most of nonpolar groups increase
their solvent accessibility, and thus the contribution of these
groups to the DCp of the protein should decrease as urea
concentration is raised; but, on the other hand, the
contribution of the backbone should increase as urea
concentration is raised as the protein is more disordered.
This latter increase should compensate for the decrease
attained by the exposition of nonpolar groups. It is also
reasonable to assume that the contribution of ionic groups
should be small, as the urea concentration is increased,
because the majority of these groups are exposed in the
folded (similar to other HPrs structures, J. L. Neira,
unpublished results) and denatured states of the protein.
In ecHPr [9] and Drosophila Notch ankyrin repeats [60]
the DCp decreased upon urea concentration. On the other
hand, in barnase (in urea) [52] and barstar (in Gdm Cl) [8],
DCp increased as the denaturant of concentration was
raised. It seems that in all the proteins studied so far, except
for barnase and scHPr, DCp always decreased with the urea
concentrations and increased with the Gdm Cl ([60] and
references therein). The increase in the DCp upon urea
concentration can be rationalized by considering that at
higher concentrations of urea, there would be more
denaturant to interact with and then the exothermic
enthalpic term superimposed to the intrinsic endothermic
term of protein unfolding would be higher. On the other
hand, in the 434 cro protein [7] and the lac-repressor DNAbinding domain [11], no denaturant-dependence of DCp was
observed. Although these differences in the behaviour of

mDCpi among the different protein models could indicate
different compensating effects between the protein–urea
interaction effects and the structural changes occurring
upon unfolding, it must be borne in mind that the most
probable origin of the discrepancies is the experimental
uncertainty in determining the third derivative of the
stability curve (mDCpi ), obtained, in most of the data, with
only one biophysical technique.


´
2178 J. L. Neira and J. Gomez (Eur. J. Biochem. 271)

Validation of the LEM in analysis of scHPr data
The use of LEM in several proteins has resulted in the
complete determination of the thermodynamic parameters
governing the folding–unfolding transition; among those
proteins are thioredoxin [4], barnase [52], barstar [8], ecHPr
[9], bsHPr [28], and the 434 Cro protein ([7] and references
therein). In the latter protein, the use of the denaturation
binding model [6,61,62] has lead to small differences in the
determination of the free energy of 0.4–0.8 kcalỈmol)1, when
compared to that obtained by the LEM. These small
differences have also been observed in other proteins when
both models have been used to determine the free energy
[39,63].
In both models, the basic statement is that denaturant
interacts preferentially with the unfolded form of the
protein. According to the binding model [6,61,62], denaturants bind stoichometrically to proteins, and the preferential interaction is given by the large number of sites upon
unfolding; however, there is not very much experimental

evidence for such a stoichometric binding. Conversely,
according to the LEM, denaturants display many weak
selective interactions with proteins, and the preferential
interaction is given by the larger solvent-accessible surface
upon unfolding. Recently, two new models have been
proposed. The first approach is based on a modelindependent approach, yielding, except at low concentrations of denaturant, similar results to those obtained by
LEM for the same proteins [13]. The second approach uses
a local-bulk partitioning model in which the urea distribution between the surface of the protein and the bulk solution
is described by a concentration-independent partition
coefficient [11,12]; the predicted free energy for the lac
repressor DNA-binding domain agrees with that obtained
by the LEM. Thus, it seems that the LEM is the most widely
used method and the other developed approaches agree well
with the results provided by the linear free-energy extrapolation.
Application of the LEM to the thermodynamic data of
scHPr accounts well for the measurement of scHPr stability,
as shown by the agreement between the chemical- (where
LEM was used to analyse the data) and thermal-unfolding
data (Fig. 5C). However, the determination of the m-value
from an independent set of measurements (i.e. the thermal
denaturation experiments at different urea concentrations)
other than those data obtained from the isothermal
chemical-denaturations (where the LEM was applied) could
provide additional validation of the use of LEM in scHPr. It
can be easily shown that a Taylor series expansion of the
free-energy up to the second term yields [13]:


DHm dTm
m1=2 ẳ

17ị
Tm dẵurea
where

midpoints were obtained yields a slope of )9.9 ± 1 KỈM)1
(data not shown). Using this slope, the enthalpies and thermal midpoints obtained by fitting the thermal CD denaturation data (Fig. 5A) yield m½ which are between 0.9 ± 0.3
(at 2 M) and 1.7 ± 0.4 (at 0 M) kcalỈmol)1ỈM)1, which agree,
within the error, with those values obtained by CD (Fig. 3C)
in the explored temperature range.
In scHPr, the m-value is temperature-dependent [either
from CD (Fig. 3C) or the fluorescence measurements
(Fig. 1C)], showing a reduction of its value as the temperature is increased (CD measurements). Similar negative
slopes in the temperature-dependence of the m-value have
been reported in the ovomucoid third domain [10] and
ribonuclease A [64]. Conversely, only in E. coli ribonuclease
H a positive slope was observed [65], and in the rest of the
proteins studied (included ecHPr [9]) the m-value was not
temperature-dependent ([60] and references therein). The
decrease in the m-value as the temperature was raised must
reflect the unfavourable entropy change and favourable
enthalpy change, associated with the accumulation of urea
in the vicinity of the protein surface which is exposed upon
unfolding (see below). Although it has been suggested that
factors other than changes in accessible surface areas play a
major role in the determination of the m-values [64],
recently, an alternative explanation to the temperaturedependence of m-values has been suggested based on the
larger exposure of accessible surface area in the folded state,
compared to the unfolded state, due to the increase in the
local structural fluctuations upon temperature [11]. These
suggestions seem to support the findings observed in

thermophylic ribonuclease H, where an absence of temperature-dependence in the m-value was observed, when
compared to the E. coli ribonuclease H (where a temperature-dependence was detected) [65], because local fluctuations in the thermophylic enzyme occurs with a lower
frequency. Based on those studies, it has been argued that
the temperature-independence of the m-value among different proteins could be due to: (a) the small mDCpi [8]; (b) to
temperature-compensating effects on the urea activity and
the binding constants to the sites exposed upon folding [28];
or (c) the experimental uncertainties. From the experiments
described in this work, it could be concluded, based
exclusively in the CD data, that the m-value in scHPr was
nearly temperature-independent (Fig. 3B). However, the
use of several techniques and several approaches to
determine the free stability curve and the repetition of the
measurements to asses the reproducibility of trends in data
and fitted parameters have allowed to conclude, within the
experimental uncertainty, the temperature-dependence of
the m-value in scHPr. Within the experimental limitations to
determine the thermodynamic parameters, we favour the
last explanation as the most probable source of the observed
temperature-independence behaviour of the m-value in
some proteins.



m1=2


@DG
ẳ
@ẵurea ẵureaẳẵurea1=2

ể FEBS 2004

and thus it can be assumed that m1=2 ffi m. The plot of the
different thermal midpoints (obtained from the thermal
denaturation experiments at different urea concentrations,
Fig. 5A) vs. the urea concentrations at which those

Thermodynamics of the interaction between scHPr
and urea
The value of mDCpi in scHPr (116 ± 4 cal mol)1ỈK)1ỈM)1,
Fig. 7) is larger than that observed in barstar [8]
(53 ± 36 calỈmol)1ỈK)1ỈM)1), where a temperature-independence behaviour of the m-value was also observed,


Ó FEBS 2004

Conformational stability of HPr of S. coelicolor (Eur. J. Biochem. 271) 2179

and also larger than those of ecHPr [9] (50 ± 20 calỈ
mol)1ỈK)1ỈM)1), Drosophila Notch ankyrin domain [60]
(70 ± 40 calỈmol)1ỈK)1ỈM)1), and a lac-repressor DNAbinding domain (< 20 calỈmol)1ỈK)1ỈM)1) (where temperature-dependence behaviours of the corresponding m-values
were observed), but similar to that observed in barnase [52]
(160 calỈmol)1ỈK)1ỈM)1, where a negative temperaturedependence of the m-value was also observed). Although
it can be argued that these discrepancies are only due to the
experimental uncertainty of the determinations, we can also
speculate that the differences might reflect either variations
in the residual structure present in each protein or larger
structural changes upon denaturant-binding.
Comparison of the other thermodynamic parameters
characterizing the binding to urea (mmi i and mmi i ) (Fig. 7)

DH
DS
with those from the unfolding of the protein (Table 2)
shows that the former are much smaller than the latter.
Similar results have also been described in other proteins
where thermodynamic magnitudes of protein-denaturant
binding have also been measured [8,9,28]. At 298 K, the
enthalpy of urea binding at 1 M of urea to scHPr is
)8.2 ± 0.4 kcalỈmol)1. The measurements obtained for
ribonuclease A and lysozyme are )8 to )10 kcalỈmol)1
[40]. In barstar [8] and ecHPr [9] the enthalpy of
interaction was less favourably ()3 kcalỈmol)1) than that
of lysozyme, but not as large as that of scHPr. It has been
argued [9] that those differences among the binding
enthalpies could be explained by the different size of the
proteins, thus the smaller proteins would have less sites to
interact with urea and the enthalpy would be less
favourable. As the number of amino acids in scHPr and
ecHPr is nearly the same, there must be other additional
factors affecting the binding to urea, which probably rely
on the exact mechanism of the urea–protein interactions
and/or the structural changes occurring upon unfolding.
To date, the exact mechanism of action of chemical
denaturants is not known, although it is clear from
structural [66] and calorimetric studies [40,52,67] that there
is an interaction (a binding) to the folded and unfolded
states of proteins, especially to certain side chains [68].
Although it must be borne in mind that the urea-binding
thermodynamic magnitudes have a large experimental
error, it is tempting to suggest that the explanation for the

behaviour of the thermodynamic parameters among the
three HPrs studied could rely on the residual structure
present in the proteins as urea concentration was
increased. Also, as a consequence of that residual
structure, it can be suggested that those differences in
the urea-binding thermodynamic parameters are associated with the presence of partially folding species. For
instance, partially folded species have been observed
during the kinetic folding pathway of ecHPr [29,30],
where a small favourable-enthalpy interaction with urea
was observed [9]; conversely, an equilibrium partially
folded species has been observed at low pH in scHPr [32],
where a larger enthalpy interaction with urea has been
observed (this work); unfortunately, ecHPr precipitated at
low pHs, which hampered the detection of any partially
folded conformation [22]. No partially folded species have
been detected in bsHPr either kinetically or at equilibrium,
but no thermodynamical denaturant-binding parameters
have been reported to date.

Conclusions
To sum up, we have shown here that the LEM can be used
to obtain the thermodynamic parameters governing the
unfolding of scHPr and its urea-binding properties. The
scHPr does not have a very high conformational stability,
although its heat-temperature of unfolding is high. Because
of the large change in heat capacity and its small enthalpy of
unfolding, a cold-denaturation has been observed in the
presence of moderate urea concentrations. Conversely to
that observed in other proteins of the same family, the
values of: (a) the change in the heat capacity upon ureabinding; and (b) the change in the enthalpy upon ureabinding are high, probably due to larger structural changes

occurring upon binding to denaturant.

Acknowledgements
We thank Fritz Titgemeyer for the generous gift of the recombinant
plasmid encoding the scHPr. We thank F. N. Barrera for critical
reading of the manuscript and for stimulating ideas. This work was
supported by Projects BIO2000-1081 (to J. G.), from the Spanish
Ministerio de Ciencia y Tecnologı´ a, CTIDIB-2002/6 (to J. L. N.) from
Generalitat Valenciana, and FIS01/0004–02 (to J. L. N.) from the
Spanish Ministerio de Sanidad y Consumo. We gratefully thank Marı´ a
´
T. Garzon, May Garcı´ a, Marı´ a C. Fuster and Javier Casanova for
excellent technical assistance. We thank the two anonymous reviewers
for their critical insights and suggestions.

References
1. Privalov, P.L. (1992) The physical basis of stability of folded
conformations of proteins. Protein Folding (Creighton, T.E., ed.),
pp. 83–126. W.H. Freeman, New York.
2. Freire, E. (1995) Differential scanning calorimetry in methods in
molecular biology. In Protein Stability and Folding: Theory and
Practice (Shirley, B.T., ed.), pp. 191–218. Humana Press, Towota,
New Jersey.
3. Schellman, J.A. (1987) The thermodynamic stability of proteins.
Annu. Rev. Biophys. Biophys. Chem. 16, 115–137.
4. Santoro, M.M. & Bolen, D.W. (1992) A test of the linear extrapolation of unfolding free energy changes over an extended
denaturant concentration range. Biochemistry 31, 4901–4907.
5. Greene, R.F. & Pace, C.N. (1974) Urea and guanidine hydrochloride denaturation of ribonuclease, lysozyme, alpha-chymotrypsin, and beta-lactoglobulin. J. Biol. Chem. 249, 5388–5393.
6. Tanford, C. (1970) Protein denaturation. C. Theoretical models
for the mechanism of denaturation. Adv. Prot. Chem. 21, 1–95.

´
7. Padmanabhan, S., Laurents, D.V., Fernandez, A.M., Elı´ as-Arnaz,
M., Ruiz-Sanz, J., Mateo, P.L., Rico, M. & Filimonov, V.V.
(1999) Thermodynamic analysis of the structural stability of phage
434 Cro protein. Biochemistry 38, 15536–15547.
8. Agashe, V.R. & Udgaonkar, J.B. (1995) Thermodynamics of
denaturation of barstar: evidence for cold denaturation and
evaluation of the interaction with guanidine hydrochloride.
Biochemistry 34, 3286–3299.
9. Nicholson, E.M. & Scholtz, J.M. (1996) Conformational stability
of the Escherichia coli HPr protein: test of the linear extrapolation
method and a thermodynamic characterization of cold denaturation. Biochemistry 35, 11369–11378.
10. Swint, L. & Robertson, A.D. (1993) Thermodynamics of unfolding for turkey ovomucoid third domain: thermal and chemical
denaturation. Protein Sci. 2, 2037–2049.
11. Felitsky, D.J. & Record, M.T. Jr (2003) Thermal and ureainduced unfolding of the marginally stable lac repressor DNA-


´
2180 J. L. Neira and J. Gomez (Eur. J. Biochem. 271)

12.

13.

14.

15.

16.

17.

18.

19.

20.

21.

22.

23.

24.

25.

26.

27.

binding domain: a model system for analysis of solute effects on
protein processes. Biochemistry 42, 2202–2217.
Courtenay, E.S., Capp, M.W., Saceker, R.M. & Record, M.T. Jr
(2000) Thermodynamic analysis of interactions between denaturants and protein surface exposed on unfolding: interpretation
of urea and guanidinium chloride m-values and their correlation
with changes in accessible surface area (ASA) using preferential
interaction coefficients and the local-bulk domain model. Proteins
41, 72–85.

´
Ibarra-Molero, B. & Sanchez-Ruiz, J.M. (1996) A model-independent, nonlinear extrapolation procedure for the characterization of protein folding energetics from solvent-denaturation
data. Biochemistry 35, 14689–14702.
Postma, P.W., Lengeler, J.W. & Jacobson, G.R. (1993) Phosphoeneolpyruvate carbohydrate phosphotransferase system of
bacteria. Microbiol. Rev. 57, 543–594.
Bruckner, R. & Titgemeyer, F. (2002) Carbon catabolite represă
sion in bacteria: choice of the carbon source and autoregulatory
limitation of sugar utilization. FEMS Microbiol. Lett. 209, 141–
148.
Meadow, N.D., Fox, D.K. & Roseman, S. (1990) The bacterial
phosphoenolpyruvate: glycose phosphotransferase system. Annu.
Rev. Biochem. 59, 497–542.
Wittekind, M., Rajagopal, P., Baranchini, B.R., Reizer, J., Saier,
M.H. Jr & Klevitt, R.E. (1992) Solution structure of the phosphocarrier protein HPr from Bacillus subtilis by two-dimensional
NMR spectroscopy. Protein Sci. 1, 1363–1376.
Van Nuland, N.A.J., Hangyi, I.W., Van Schaik, R.C., Berendsen,
H.J.C., Van Gusteren, W.F., Scheek, R.M. & Robillard, G.T.
(1994) The high-resolution structure of the histidine-containing
phosphocarrier protein HPr from Escherichia coli determined by
restrained molecular dynamics from nuclear magnetic resonance
nuclear Overhauser effect data. J. Mol. Biol. 237, 544–559.
Maurer, T., Doker, R., Gorler, A., Hengstenberg, W. & Kalbitzer,
H.R. (2001) Three-dimensional structure of the histidine-containing phosphocarrier protein (HPr) from Enterococcus faecalis
in solution. Eur. J. Biochem. 268, 635–644.
Herzberg, O., Reddy, P., Sutrin, S., Saier, M.H. Jr, Reizer, J. &
Kapafia, G. (1992) Structure of the histidine-containing phos˚
phocarrier protein HPr from Bacillus subtilis at 2.0- A resolution.
Proc. Natl Acad. Sci. USA 89, 2499–2503.
Jia, Z., Quail, J.W., Waygood, E.B. & Delbaere, L.T.J. (1993) The
˚

2.0-A resolution structure of Escherichia coli histidine-containing
phosphocarrier protein HPr: a redetermination. J. Biol. Chem.
268, 22490–22501.
van Nuland, N.A.J., Meijberg, W., Warner, J., Forge, V., Scheek,
R.M., Robillard, G.T. & Dobson, C.M. (1998) Slow cooperative
folding of a small globular protein HPr. Biochemistry 37, 622–637.
Pastore, A., Saudek, V., Ramponi, G. & Williams, R.J.P. (1992)
Three-dimensional structure of acylphosphatase. Refinement and
structure analysis. J. Mol. Biol. 224, 427–440.
Bentley, S.D., Chater, K.F. et al. & Hoopwood, D.A. (2002)
Complete genome sequence of the model actinomycete Streptomyces coelicolor A3 (2). Nature 417, 141–147.
Titgemeyer, F., Walkenhorst, J., Cui, X., Reizer, J. & Saier, M.H.
Jr (1994) Proteins of the phosphoenolpyruvate: sugar phosphotransferase system in Streptomyces: possible involvement in the
regulation of antibiotic production. Res. Microbiol. 145, 89–92.
Titgemeyer, F., Walkenhorst, J., Reizer, J., Stuiver, M.H., Cui, X.
& Saier, M.H. Jr (1995) Identification and characterization of
phosphoenolpyruvate: fructose phosphotransferase systems in
three Streptomyces species. Micorobiology 141, 51–58.
Parche, S., Schmid, R. & Titgemeyer, F. (1999) The PTS system of
Streptomyces coelicolor: identification and biochemical analysis of
a histidine phosphocarrier protein HPr encoded by the gene ptsH.
Eur. J. Biochem. 265, 308–317.

Ó FEBS 2004
28. Scholtz, J.M. (1995) Conformational stability of HPr: the histidine-containing phosphocarrier protein from Bacillus subtilis.
Protein Sci. 4, 35–43.
29. Azuaga, A.I., Canet, D., Smeenk, G., Berends, R., Titgemeyer, F.,
Duurkens, R., Mateo, P.L., Scheek, R.M., Robillard, G.T.,
Dobson, C.M. & van Nuland, N.A.J. (2003) Characterization of
single-tryptophan mutants of histidine-containing phosphocarrier

protein: evidence for local rearrangements during folding from
high concentrations of denaturant. Biochemistry 42, 4883–4895.
30. Canet, D., Lyon, C.E., Scheek, R.M., Robillard, G.T., Dobson,
C.M., Hore, P.J. & van Nuland, N.A.J. (2003) Rapid formation of
non-native contacts during the folding of HPr revealed by realtime photo-CIDNP NMR and stopped-flow fluorescence experiments. J. Mol. Biol. 330, 397–407.
31. Gunasekaran, K., Eyles, S.J., Haggler, A.T. & Gierasch, L.M.
(2001) Keeping it in the family: folding studies of related proteins.
Cur. Opin. Struct. Biol. 11, 83–93.
´
´
32. Fernandez-Ballester, G., Maya, J., Martı´ n, A., Parche, S., Gomez,
J., Titgemeyer, F. & Neira, J.L. (2003) The histidine-phosphocarrier protein of Streptomyces coelicolor folds by a partially folded species at low pH. Eur. J. Biochem. 270, 2254–2267.
33. Pace, C.N. (1986) Determination and analysis of urea and
guanidine hydrochloride denaturation curves. Methods Enzymol.
131, 266–280.
34. Pace, C.N. & Scholtz, J.M. (1997) Measuring the conformational
stability of a protein. In Protein Structure (Creighton, T.E., ed.),
2nd edn, pp. 253–259. Oxford University Press, Oxford.
35. Privalov, P.L. (1979) Stability of proteins: small globular proteins.
Adv. Protein Chem. 33, 167–241.
36. Chen, B.L. & Schellman, J.A. (1989) Low-temperature unfolding
of a mutant of phage T4 lysozyme. 1. Equilibrium studies.
Biochemistry 28, 685–691.
37. Becktel, W.J. & Schellman, J.A. (1987) Protein stability curves.
Biopolymers 26, 1859–1877.
38. Vuilleumier, S., Sancho, J., Lowenthal, R. & Fersht. A.R. (1993)
Circular dichroism studies of barnase and its mutants: characterization of the contribution of aromatic side chains. Biochemistry
32, 10303–10313.
39. Pace, C.N. & Laurents, D.V. (1989) A new method for
determining the heat capacity change for protein folding.

Biochemistry 28, 2520–2525.
40. Makhatadze, G.I. & Privalov, P.L. (1992) Protein interactions
with urea and Gdm chloride. A calorimetric study. J. Mol. Biol.
226, 491–505.
41. Cornish-Bowden, A. (2002) Enthalpy-entropy compensation: a
phantom phenomenon. J. Biosci. 27, 121–126.
42. Sharp, K. (2001) Enthalpy-entropy compensation: Fact or artefact. Prot. Sci. 10, 661–667.
43. Huang, G.S. & Oas, T.G. (1996) Heat and cold denatured states of
monomeric lambda repressor are thermodynamically and conformationally equivalent. Biochemistry 35, 6173–6180.
44. DeKoster, G.T. & Robertson, A.D. (1995) Cold denaturation of
CheY. J. Mol. Biol. 249, 529–534.
45. Chi, Y.-H., Kumar, T.K., Wang, H.-M., Ho, M.-C., Chiu, I.-M.
& C. (2001) Thermodynamic characterization of the human
acidic fibroblast growth factor: evidence for cold denaturation.
Biochemistry 40, 7746–7753.
46. Rees, D.C. & Robertson, A.D. (2001) Some thermodynamic
implications for the thermostability of proteins. Protein Sci. 10,
1187–1194.
47. Kumar, S., Tsai, C.J. & Nussinov, R. (2002) Maximal stabilities of
reversible two-state proteins. Biochemistry 41, 5359–5374.
48. Robertson, A.D. & Murphy, K.P. (1997) Protein Structure and
the Energetics of Protein Stability. Chem. Rev. 97, 1251–1267.
49. Privalov, P.L. (1990) Cold denaturation of proteins. Crit. Rev.
Biochem. Mol. Biol. 25, 281–305.


Ó FEBS 2004

Conformational stability of HPr of S. coelicolor (Eur. J. Biochem. 271) 2181

50. Tsai, C.J., Maizel, J.V. & Nussinov, R. (2002) The hydrophobic
effect. A new insight from cold-denaturation and two-state water
structure. Crit. Rev. Biochem. Mol. Biol. 37, 55–69.
51. Ruelle, P. & Kesselring, U.W. (1998) The hydrophobic effect. 1. A
consequence of the mobile order in H-bonded liquids. J. Pharm.
Sci. 87, 987–997.
52. Johnson, C.M. & Fersht, A.R. (1995) Protein stability as a function of denaturant concentration: the thermal stability of barnase
in the presence of urea. Biochemistry 34, 6795–6804.
53. Privalov, P.L. & Makhatadze, G.I. (1990) Heat capacity of proteins. II. Partial molar heat capacity of the unfolded polypeptide
chain of proteins: protein unfolding effects. J. Mol. Biol. 213,
385–391.
54. Privalov, P.L., Tiktopulo, E.I., Venyaminov, S.Y., Griko, Y.V.,
Makhatadze, G.I. & Khechninashvili, N.N. (1989) Heat capacity
and conformation of proteins in the denatured state. J. Mol. Biol.
205, 737–750.
55. Zou, Q., Habermann-Rottinghaus, S.M. & Murphy, K.P. (1998)
Urea effects on protein stability: hydrogen bonding and the
hydrophobic effect. Proteins 31, 107–115.
56. Kresheck, G.C. & Benjamin, L. (1964) Calorimetric studies of
the hydrophobic nature of several protein constituents and
ovoalbumin in water an aqueous urea. J. Phys Chem. 68, 2476–
2486.
57. de Visser, C., Perron, G. & Desnoyers, J.E. (1977) Volumes and
heat capacities of ternary aqueous systems at 25 °C. Mixtures of
urea, tert-butyl alcohol, dimethylformamide and water. J. Am.
Chem. Soc. 98, 5894–5900.
58. Enea, O. & Jollcoeur, C. (1982) Heat capacity and Volumes of
several oligopeptides in urea-water mixtures at 25 °C. Some
implications for protein unfolding. J. Phys. Chem. 86, 3870–
3881.

59. Hakin, A.W. & Hedgiw, G.R. (2001) Group additivity calculations of the thermodynamic properties of unfolded proteins in
aqueous solution: a critical comparison of peptide-based and
HKF models. Biophys. Chem. 89, 253–264.
60. Zweifel, M.E. & Barrick, D. (2002) Relationships between the
temperature dependence of solvent denaturation and the denaturant dependence of protein stability curves. Biophys. Chem.
101–102, 221–237.
61. Schellman, J.A. (1987) Selective binding and solvent denaturation.
Biopolymers 26, 549–559.
62. Aune, K.C. & Tandford, C. (1969) Thermodynamics of the
denaturation of lysozyme by guanidine hydrochloride. I.
Depdendence on pH at 25 degrees. Biochemistry 8, 4586–4590.
63. Filimonov, V.V., Azuaga, A.I., Viguera, A.R., Serrano, L. &
Mateo, P.L. (1999) A thermodynamic analysis of a family of small
globular proteins: SH3 domains. Biophys. Chem. 77, 195–208.
64. De Koster, G.T. & Robertson, A.D. (1997) Calorimetricallyderived parameters for protein interactions with urea and
guanidine-HCl are not consistent with denaturant m-values.
Biophys. Chem. 64, 59–68.
65. Hollien, S. & Marqusee, S. (1999) A thermodynamic comparison
of mesophilic and thermophilic ribonucleases H. Biochemistry 38,
3831–3836.
66. Pike, A.C.W. & Acharya, K.R. (1994) A structural basis for the
interaction of urea with lysozyme. Protein Sci. 3, 706–710.
67. Zolkiewski, M., Nosworthy, N.J. & Ginsburg, A. (1995) Ureainduced dissociation and unfolding of dodecameric glutamine
synthetase from Escherichia coli: calorimetric and spectral studies.
Protein Sci. 3, 706–710.
68. Tirado-Rives, J., Orozco, M. & Jorgensen, W.L. (1997) Molecular
dynamics simulations of the unfolding of barnase in water and 8
M aqueous urea. Biochemistry 36, 7313–7329.