Báo cáo khoa học: The histidine-phosphocarrier protein of Streptomyces coelicolor folds by a partially folded species at low pH ppt

The histidine-phosphocarrier protein of
Streptomyces coelicolor
folds by a partially folded species at low pH
Gregorio Ferna
´
ndez-Ballester
1
, Javier Maya
1
, Alejandro Martı
´
n
1
, Stephan Parche
2,
*, Javier Go
´
mez
1
,
Fritz Titgemeyer
2
and Jose
´
L. Neira
1,3
1
Instituto de Biologı
´
a Molecular y Celular, Universidad Miguel Herna
´

ndez, Elche (Alicante), Spain;
2
Lehrstuhl fu
¨
r Mikrobiologie,
Friedrich-Alexander-Universita
¨
t Erlangen-Nu
¨
rnberg, Germany;
3
Instituto de Biocomputacion y Fisica de los sistemas complejos,
Zaragoza, Spain
The folding of a 93-residue protein, the histidine-phospho-
carrier protein of Streptomyces coelicolor,HPr,hasbeen
studied using several biophysical techniques, namely fluo-
rescence, 8-anilinonaphthalene-1-sulfate binding, circular
dichroism, Fourier transform infrared spectroscopy, gel
filtration chromatography and differential scanning calori-
metry. The chemical-denaturation behaviour of HPr, fol-
lowed by fluorescence, CD and gel filtration, at pH 7.5 and
25 °C, is described as a two-state process, which does not
involve the accumulation of thermodynamically stable
intermediates. Its conformational stability under those con-
ditions is DG ¼ 4.0 ± 0.2 kcalÆmol
)1
(1 kcal ¼ 4.18 kJ),
which makes the HPr from S. coelicolor the most unstable
member of the HPr family described so far. The stability of
the protein does not change significantly from pH 7–9, as

concluded from the differential scanning calorimetry and
thermal CD experiments. Conformational studies at low pH
(pH 2.5–4) suggest that, in the absence of cosmotropic
agents, HPr does not unfold completely; rather, it accumu-
lates partially folded species. The transition from those
species to other states with native-like secondary and tertiary
structure, occurs with a pK
a
¼ 3.3 ± 0.3, as measured by
the averaged measurements obtained by CD and fluores-
cence. However, this transition does not agree either with: (a)
that measured by burial of hydrophobic patches (8-anilino-
naphthalene-1-sulfate binding experiments); or (b) that
measured by acquisition of native-like compactness (gel-fil-
tration studies). It seems that acquisition of native-like
features occurs in a wide pH range and it cannot be
ascribed to a unique side-chain titration. These series of
intermediates have not been reported previously in any
member of the HPr family.
Keywords: folding; molten-globule; protein stability; PTS;
structure.
The phosphoenolpyruvate phosphotransferase system
(PTS) catalyzes the uptake and phosphorylation of carbo-
hydrates in most bacterial species, via a cascade of several
proteins [1]. Enzyme I (EI), the first protein in the cascade,
is autophosphorylated by phosphoenolpyruvate, yielding
phosphorylated EI (P-EI). P-EI acts as a phosphoryl donor
to the histidine-phosphocarrier protein (HPr). Phosphoryl-
ated HPr, in turn, donates the phosphoryl moiety to a group
of specific sugar-transporter proteins, known as enzymes II

(EII). The transfer of the phosphoryl moiety from EII to the
sugar occurs concomitantly to its transport into the cell.
HPr, the smallest protein in the cascade, is thought to be the
key component in that cascade, because it phosphorylates
all sugar-specific EII proteins. EI phosphorylates HPr at
the imidazole ring of the highly conserved histidine [1]. This
phosphorylation is common to both Gram-positive and
Gram-negative bacteria. However, in Gram-positive bac-
teria there is also an additional regulatory phosphorylation
site in HPr at the conserved Ser46. This regulatory site is
thought to be involved in carbon catabolite repression of
several genes, and as a transcription regulator of several
operons in Gram-positive bacteria [2].
Streptomyces are soil-dwelling, high GC Gram-positive
actinomycetes which grow on a variety of carbon sources,
such as cellulose and many monosaccharides and disaccha-
rides. They are the source of approximately two-thirds of all
natural antibiotics currently produced by the pharmaceu-
tical industry. Despite their importance, our knowledge on
nutrient sensing, carbohydrate transport and regulation is
very poor. The complete sequence of Streptomyces coeli-
color has been sequenced, showing the largest number of
genes found in any bacteria [3]. The presence of the different
components of the PTS in S. coelicolor has been reported,
and the corresponding proteins cloned and expressed [4–9].
EI and HPr proteins from S. coelicolor are similar, in
Correspondence to J. L. Neira, Instituto de Biologı
´
a Molecular y
Celular, Edificio Torregaita

´
n, Universidad Miguel Herna
´
ndez,
Avda del Ferrocarril s/n, 03202, Elche (Alicante), Spain.
Fax: +34 966658758, Tel.: +34 966658459,
E-mail:
Abbreviations: ANS, 8-anilinonaphthalene-1-sulfonate ammonium
salt; DSC, differential scanning calorimetry; DC
p
, heat capacity
change; DH
cal
, the calorimetric enthalpy change; DH
VH
, the van’t Hoff
enthalpy change; EI, enzyme I; EII, enzyme II; FTIR, Fourier trans-
form infrared spectroscopy; GdmCl, guanidine hydrochloride; HPr,
histidine-phosphocarrier protein; PTS, the phosphoenolpyruvate-
dependent sugar phosphotransferase system; T
m
,thermal
denaturation midpoint.
*Present address:Nestle
´
Research Center, Vers-chez-les-Blanc,
CH-100, Lausanne 26, Switzerland.
(Received 7 January 2003, revised 5 March 2003,
accepted 26 March 2003)
Eur. J. Biochem. 270, 2254–2267 (2003) Ó FEBS 2003 doi:10.1046/j.1432-1033.2003.03594.x

molecular size and activity, to those corresponding proteins
of other microorganisms previously characterized [4].
Because of its key role in the PTS and its small size, we
are studying HPr from S. coelicolor as a model for:
(a) advancing our knowledge of the PTS enzymatic activity,
and then, its involvement in antibiotic production; and
(b) research in protein folding.
The structures of HPrs from several species have been
studied by NMR spectroscopy [10–12 and references therein]
and X-ray diffraction [13,14,12 and references therein]. HPrs
from those species show a classical open-face b-sandwich
fold consisting of three a helices packed against a four-
stranded antiparallel b sheet; this fold has been related to
other proteins with no apparent relationship in function,
such as ferredoxin and diphosphate kinase [15,16]. HPr
from S. coelicolor contains 93 amino acid residues; it lacks
cysteine and tyrosine, but it has a large number of alanine,
valine and leucine residues. Furthermore, the protein only
contains one tryptophan residue and one phenylalanine
residue, which makes it a good model to follow its folding
mechanism and other biochemical features by using fluor-
escence spectroscopy. Assignment and preliminary NMR
studies of the HPr from S. coelicolor indicate that its
structure is similar to that observed in other members of the
HPr family (J. L. Neira, unpublished results). As the HPr
from S. coelicolor has a similar structure, but a completely
different amino acid sequence to those HPrs from Escheri-
chia coli or Bacillus subtilis, whose structure and folding
properties have been described, it is important to understand
if the structure, the sequence or both determine the

conformational stability and biochemical properties in the
HPr family. There is much current interest in determining
the extent to which related proteins share stability and
folding features [17]. The exploration of the folding and
stability among the different HPr members will allow us to
decide whether there is a common thermodynamic equili-
brium behaviour in this important family.
In this study, we use several biophysical techniques (CD,
fluorescence, 8-anilinonaphthalene-1-sulfate (ANS) binding,
differential scanning calorimetry (DSC), thermal CD, FTIR
and gel filtration chromatography) to follow the folding of
HPr from S. coelicolor. Our findings indicate that the folding
of HPr can be adequately described as a two-state process
without the accumulation of intermediates at neutral and
moderately basic conditions (pH 7–9) at 25 °C. The stability
of the protein, at pH 7.5 and 25 °C, as obtained by
chemical and thermal denaturation experiments, is low:
DG ¼ 4.0±0.3kcalÆmol
)1
. At moderately acid pH values
(pH 2.5–4), in the absence of cosmotropic agents, the protein
undergoes noncooperative thermal denaturations and it
accumulates partially folded species. Although extensive
folding studies have been carried out with the E. coli HPr
[16], no such species have been previously observed nor
characterized. Thus, this is the first member of the HPr where
partially folded species have been found.
Experimental procedures
Materials
Urea and guanidine hydrochloride (GdmCl) ultra-pure

were from ICN Biochemicals. Urea and GdmCl molecular
biology grade, imidazole, trizma base, NaCl and ANS were
from Sigma. 2-Mercaptoethanol was from Bio-Rad, and
the Ni
2+
-resin was from Invitrogen. Glutaraldehyde was
from Fluka. Dialysis tubing was from Spectrapore with a
molecular mass cut-off of 3500 Da. Standard suppliers were
used for all other chemicals. Water was deionized and
purified on a Millipore system. Urea and GdmCl stock
solutions were prepared gravimetrically and filtered using
0.22-lm syringe driven filters from Millipore. Exact con-
centrations of urea stock solutions were calculated from
the refractive index of the solution using an Abbe 325
refractometer [18].
Protein expression and purification
The HPr clone comprises residues 1–93 (with the extra
methionine at the N terminus) and the His
6
-tag at the
N terminus. We have carried out all the studies with this
construction as its structure, as observed by NMR (unpub-
lished results), is similar to that found in other members of
the HPr family and the His
6
-tag is disordered in solution,
making no contacts with the rest of the protein. Recom-
binant protein was expressed in E. coli C43 strain [19], and
purified using Ni
2+

-chromatography. To eliminate any
protein or DNA bound to the resin, coeluting with the
protein, an additional gel filtration chromatography step
was carried out by using a Superdex 75 16/60 gel filtration
column, running on an AKTA-FPLC (Amersham Bio-
sciences) system. The yield was 25–30 mg of protein per litre
of culture. Protein was more than 99% pure as judged
by SDS protein-denaturing gels. Also, mass spectrometry
analysis was carried out in a MALDI-TOF instrument; only
one peak was observed. The samples were dialyzed exten-
sively against water and stored at )80 °C. Samples were
prepared by dissolving the lyophilized protein in deionized
water (unfolding) or in 8
M
urea (in the refolding experi-
ments) and adding the proper buffer solution. Protein
concentration was calculated from the absorbance of stock
solutions measured at 280 nm, using the extinction coeffi-
cients of model compounds [20].
Protein without the His
6
-tag was obtained using the
thrombin-cleavage capture kit from Novagen (Germany).
This protein was used as a control experiment to address the
importance of the His
6
-tag in the biophysical properties of
the protein under different pH conditions.
Cross-linking experiments
Cross-linking reactions were performed at 25 °Cinthe

corresponding buffers at different protein concentrations by
addition of glutaraldehyde to a final concentration of 4 m
M
.
Reactions were stopped after 15 min by addition of SDS-
buffer.
Fluorescence measurements
All fluorescence spectra were collected on a SLM 8000
spectrofluorometer (Spectronics Instruments, Urbana, IL),
interfaced with a Haake water bath, at 25 °C. All measure-
ments were corrected for wavelength dependence on the
exciting-light intensity through the use of the quantum
counter rhodamine B in the reference channel [21]. Sample
Ó FEBS 2003 Stability of Streptomyces coelicolor HPr (Eur. J. Biochem. 270) 2255
concentration was in the range 6–20 l
M
, and the final
concentration of the buffer was, in all cases, 10 m
M
.A
0.5-cm path-length quartz cell (Hellma) was used.
Steady state fluorescence measurements. All protein sam-
ples were excited at 280 nm, as excitation at 295 nm yielded
the same spectrum with smaller intensity (data not shown).
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 increment of wavelength
was set to 1 nm. Blank corrections were made in all spectra.
The urea titrations, followed either by fluorescence or

CD, were carried out by two different procedures: (a)
dilution of the proper amount of the 8
M
denaturant stock
solution and leaving the samples at 25 °C, for at least 8 h
prior to performing the experiments; or (b) by directly
titrating the protein with urea. No differences were observed
between the procedures. As the concentration of urea was
increased, the fluorescence spectra were red-shifted and their
intensities decreased (data not shown). Experiments carried
out at different protein concentrations (6–20 l
M
) did not
show any difference in the thermodynamic parameters
obtained (data not shown).
In the pH-induced unfolding experiments, followed either
by fluorescence or CD, the pH was measured after comple-
tion of the experiments with an ultra-thin Aldrich electrode in
a Radiometer (Copenhagen) pH meter. The pH range
explored using both techniques was 1.5–12. The buffers were:
pH 1.5–3.0, phosphoric acid; pH 3.0–4.0, formic acid;
pH 4.0–5.5, acetic acid; pH 6.0–7.0, NaH
2
PO
4
;pH7.5–
9.0, Tris acid; pH 9.5–11.0, Na
2
CO
3

; pH 11.5–12, Na
3
PO
4
.
Fluorescence quenching experiments. Quenching of intrin-
sic tryptophan fluorescence by iodide or acrylamide [21] was
examined at different pH values. Excitation was at 280 nm,
and emission was measured from 300 to 400 nm. In
experiments employing KI as a quencher, ionic strength
was kept constant by addition of KCl; also, Na
2
S
2
O
3
was
added to a final concentration of 0.1
M
to avoid formation
of I
3
–
. The slit width was set at 8 nm for both excitation and
emission. The dynamic and static quenching constants for
acrylamide were obtained by fitting the data from different
wavelengths (in the range 330–340 nm) to the Stern–Volmer
equation, which includes an exponential term to account for
static quenching [21]:
F

0
F
¼ 1 þ K
sv
½Xe
ðv½XÞ
ð1Þ
where K
sv
is the Stern–Volmer constant for collisional
quenching and v is the static quenching constant. Iodide
quenching did not show a significant static component, and
then the exponential term was not included in the fitting of
Eqn (1). The range of concentrations used in both quenc-
hers was 0–0.7
M
. Experiments carried out at different
protein concentrations did not show any difference in the
parameters obtained (data not shown).
ANS binding. ANS binding was measured by collecting
fluorescence spectra at different pH values in the presence of
50 l
M
dye. Excitation wavelength was 380 nm, and emis-
sion was measured from 400 to 600 nm. Slit widths were
4 nm for excitation and 8 nm for emission. Stock solutions
ofANSwerepreparedinwateranddilutedintothesamples
to the above final concentration. In all cases, the blank
solutions were subtracted from the corresponding spectra.
Experiments carried out a different protein concentrations

did not show any difference (data not shown).
Circular dichroism
Circular dichroism 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 periodically calibrated with (+)10-cam-
phorsulfonic acid. Isothermal wavelength spectra at differ-
ent pH values were acquired with a scan speed of
50 nmÆmin
)1
, and a response time of 2 s and averaged over
four scans at 25 °C. Far-UV measurements were performed
using 14–295 l
M
of protein in 10 m
M
buffer, using 0.1- or
0.2-cm pathlength cells (Hellma). During the pH titration
experiments no significant changes either in the shape or in
the molar ellipticity were observed as the concentration of
protein was increased; thus, we can rule out the presence
of concentration-dependence at those pH values. Near-UV
spectra were acquired using 30–40 l
M
of protein in a 0.5-cm
pathlength cell (Hellma). All spectra were corrected by
subtracting the proper baseline. To allow for comparison at
different pH values and different urea concentrations, raw
ellipticity was converted to molar ellipticity [22].
Thermal-denaturation experiments were performed at

constant heating rates of 60 °CÆh
)1
and 30 °CÆh
)1
;the
response time was 8 s. Thermal scans were collected in the
far-UV region at 222 nm from 25 °C(or5°C) to 90 °C(or
95 °C) in 0.1-cm pathlength cells (Hellma) with a total
protein concentration of 40–100 l
M
. Conditions were the
same as those reported in the steady-state far-UV experi-
ments. The reversibility of thermal transitions was tested by
recording a new scan after cooling down to 5 °Cthe
thermally denatured samples. To check also for reversibility,
we carried out the reheating experiments at different speeds
to the heating measurements; no differences among the
scans acquired at different speeds were observed at those pH
values where HPr unfolds reversibly. Every thermal dena-
turation experiment was repeated at least three times with
fresh new samples at different concentrations. The measured
thermodynamic parameters did not change when experi-
ments were acquired at different protein concentrations. In
all cases, after the reheating experiment, the samples were
transparent and no precipitation was observed. The possi-
bility 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
the scans.
In the urea-denaturation experiments, far-UV CD spec-

tra were acquired at a scan speed of 50 nmÆmin
)1
and four
scans were recorded and averaged at 25 °C. The response
time was 2 s. The pathlength cell was 0.1 cm, with protein
ranging in 10–30 l
M
. Spectra were corrected by subtracting
the proper baseline in all cases. The chemical denaturation
reaction was fully reversible, as demonstrated by the
agreement between the folding and unfolding curves (data
2256 G. Ferna
´
ndez-Ballester et al.(Eur. J. Biochem. 270) Ó FEBS 2003
not shown). Each chemical denaturation experiment was
repeated at least three times with new samples. Experiments
carried out at different protein concentrations did not show
any difference.
Analysis of the pH- and chemical-denaturation curves,
and free energy determination
The average energy of emission (or the intensity weighted
average of the inverse wavelengths) in the fluorescence
spectra, <k>,wascalculatedasdefinedin[23].
The pH-denaturation experiments were analyzed assu-
ming that both species, protonated and deprotonated,
contributed to the fluorescence (or CD) spectrum:
X ¼
ðX
a
þ X

b
10
ðpHÀpK
a
Þ
Þ
ð1 þ 10
ðpHÀpK
a
Þ
Þ
ð2Þ
where X is the physical property being observed (ellipticity
or fluorescence), X
a
is the physical property being observed
at low pH values (that is, the fluorescence or ellipticity of the
acid form), X
b
is the physical property observed at high pH
values, and pK
a
is the apparent pK of the titrating group.
The apparent pK
a
reported was obtained from three
different measurements, prepared with new samples. In
the fluorescence experiments, the determinations were
carried out using either the <k> or the maximum
wavelength in fitting the Eqn (2). In the CD experiments,

the ellipticity at 222 nm was the chosen parameter, either in
the pH-denaturation or chemical-denaturation experiments.
To facilitate comparison among the different biophysical
techniques, data were converted to the fraction of folded
and unfolded molecules [24].
The denaturation data obtained by fluorescence or CD
were fit to the two-state equation:
X ¼
ðX
N
þ X
D
e
ðÀDG=RTÞ
Þ
ð1 þ e
ðÀDG=RTÞ
Þ
ð3Þ
where X
N
and X
D
are the corresponding fractions of the
folded (N) and unfolded states (U), respectively, which were
allowed to change linearly with either the denaturant
(X
N
¼ a
N

+ b
N
[D], and X
D
¼ a
D
+ b
D
[D]), or the tem-
perature (that is, X
N
¼ a
N
+ b
N
T and X
D
¼ a
D
+ b
D
T ),
R is the gas constant and T is the temperature in K.
Chemical-denaturation curves were analyzed using a
two-state unfolding mechanism, according to the linear
extrapolation model: DG ¼ m([D]
50%
) [D]) [20], where DG
is the free energy of denaturation, and [D]
50%

is the denat-
urant concentration at the midpoint of the transition. The
chemical-denaturation-binding model [25,26] was also used
for the fitting of the chemical denaturation data, but no
reliable parameters were obtained either for DG or m (data
not shown). Thus, the linear extrapolation method was used
in all the conformational stability calculations.
The change in free energy upon temperature in Eqn (3) is
given by the Gibbs–Helmholtz equation:
DGðT Þ¼DH
VH
1 À
T
T
m

À DC
p
ðT
m
À TÞþT ln
T
T
m

ð4Þ
where DH
VH
is the van’t Hoff enthalpy change. By
substitution of this expression in Eqn (3), we obtain DH

VH
,
T
m
and DC
p
of the thermal experiments.
Fittings by nonlinear least-squares analysis to Eqns (1, 2
and 3) were carried out by using the general curve fit option
of
KALEIDAGRAPH
(Abelbeck software).
Gel filtration chromatography
Analytical gel filtration experiments were carried out by
using an analytical gel filtration Superdex 75 HR 16/60
(Amersham Biosciences) running on an AKTA FPLC
system at 25 °C. Flow rates of 0.8 mLÆmin
)1
(at high urea
concentrations) or 1 mLÆmin
)1
were used. The elution
buffers for the pH experiments were those described above
with 150 m
M
NaCl added to avoid non-specific inter-
actions with the column. To check for the presence of
aggregated species at low pH values, protein concentra-
tions ranged from 20 to 60 l
M

. No differences in the
elution volumes were observed among the different
concentrations used.
The chemical denaturation experiments were acquired at
pH 7.5, 10 m
M
phosphate buffer and 150 m
M
NaCl.
Protein concentration was 20–60 l
M
and absorbance was
monitored at 280 nm. No differences in the elution volumes
were observed when the protein concentration was
increased. The column was calibrated using the gel filtration
low relative molecular mass calibration kit (Amersham
Biosciences). The standards used and their corresponding
Stokes radii were: ribonuclease A (16.4 A
˚
); chymotrypsi-
nogen (20.9 A
˚
); ovoalbumin (30.5 A
˚
), and bovine serum
albumin (35.5 A
˚
)[27].
The elution of a macromolecule in gel filtration experi-
ments is usually given by the partition coefficient, r,which

is defined as the fraction of solvent volume within the gel
matrix accessible to the macromolecule [28]. The r of
protein standards and HPr were calculated by:
r ¼
ðV
e
À V
o
Þ
V
i
ð5Þ
where V
e
is the elution volume of the protein, and, V
o
and V
i
are the void and internal volumes of the column, respect-
ively. The values of those volumes are, respectively,
8.13 ± 0.06 mL and 28.43 ± 0.03 mL. The V
o
and V
i
volumes were, respectively, determined using Blue dextran
(5 mgÆmL
)1
,in10m
M
phosphate buffer plus 150 m

M
NaCl) and
L
-tryptophan (0.5 mgÆmL
)1
, in the above buffer)
by averaging four measurements for each agent.
There is a linear relationship between the molecular
Stokes radius, R
s
, and the inverse error function comple-
ment of r (erfc
)1
(r)), given by [28,29]:
R
s
¼ a þ bðerfc
À1
ðrÞÞ ð6Þ
where a and b are the calibration constants for the column.
Fitting of the calculated erfc
)1
(r) to the above equation by
linear least-squares analysis was carried out with
KALEIDA-
GRAPH
(Abelbeck software) working on a PC computer.
Once the calibration parameters are obtained, the Stokes
radius of any macromolecule can be obtained by using
Eqn (6).

Ó FEBS 2003 Stability of Streptomyces coelicolor HPr (Eur. J. Biochem. 270) 2257
Fourier transform infrared spectroscopy
The protein was lyophilized and dissolved in deuterated
buffer at different pH values. The buffer was composed of
0.1
M
NaCl, 0.1 m
M
ethylenediaminetetracetate, 0.02%
NaN
3
,10m
M
sodium acetate, 10 m
M
N-(1-morpholino)-
propane-sulfonic acid, and 10 m
M
3-(cyclohexylamino)-
1-propane-sulfonic acid. No corrections were done for the
isotope effects in the measured pH. Samples of HPr at a
final concentration of 5–6 mgÆmL
)1
were placed between a
pair of CaF
2
windows separated by a 50-lm thick spacer
in a Harrick Ossining demountable cell. Spectra were
acquired on a Nicolet 520 instrument equipped with a
deuterated triglycine sulfate detector and thermostated

with a Braun water bath at 25 °C. The cell container was
continuously filled with dry air. Usually 600 scans per
sample were taken, averaged, apodized with a Happ–
Genzel function, and Fourier transformed to give a final
resolution of 2 cm
)1
. The contributions of buffer spectra
were subtracted, and the resulting spectra used for analysis
after smoothing. The spectra smoothing was carried out
by using the maximum entropy method [30]. Derivation of
FTIR spectra was performed using a power of 3 and a
breakpoint of 0.3. Fourier self-deconvolution was per-
formed using a Lorentzian bandwidth of 18 cm
)1
and a
resolution enhancement factor of 2 [30]. The prediction of
protein secondary structure was quantified by deconvolu-
tion of the amide I band, as described elsewhere [31],
yielded essentially the same percentages of a helix, bturns
and b sheet, which have been observed in the NMR
structure (unpublished results).
Thermal denaturation experiments followed by FTIR
were performed at a protein concentration of 6 mgÆmL
)1
,
with a scanning rate of 50 °CÆh
)1
, and acquired every 5 °C.
Differential scanning calorimetry
DSC experiments [32] 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 dis-
solved in 10 m
M
phosphate buffer, pH 7.5 and dialyzed
extensively against 2 L of the same buffer (twice) at 4 °C.
Protein concentration was calculated from the absorbance
of the solution at 280 nm [20]. Samples were degassed
under vacuum for 10–15 min with gentle stirring prior to
being loaded into the calorimetric cell. DSC experiments
were performed under a constant external pressure of
1 bar in order to avoid bubble formation, and samples
were heated at a constant scan rate of 60 °CÆh
)1
. Once the
first scan was completed, the samples were cooled in situ
downto10°C for 40 min and rescanned under the same
experimental conditions in order to check the reversibility
of the heat-induced denaturation reaction. 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 analyzed using the software
package
ORIGIN
(Microcal Software, Inc.).

Results
pH-induced unfolding of HPr
To examine how the secondary and tertiary structure of
HPr changes with the pH, we used multiple spectroscopic
techniques, which give complementary information about
the melting of the secondary and/or tertiary structure of
the protein.
Fluorescence experiments. We used fluorescence to map
any change in the tertiary structure of the protein upon pH
changes [33]. HPr has one tryptophan residue, which is at
the C-terminal region of the first b strand. The emission
fluorescence spectrum of native HPr showed a maximum at
337 nm at neutral pH (Fig. 1A), indicating Trp burial. As
the pH decreased, the maxima wavelength were red-shifted
towards 350 nm, and the fluorescence intensities increased.
Fig. 1. pH-induced unfolding of HPr followed by fluorescence and ANS
binding. (A) Steady-state fluorescence: the average energy of emission
(filled circles) and the maxima wavelength (open circles) are repre-
sented vs. the pH. (B) ANS-binding experiments: the average energy
(filled circles) and the maxima wavelength (open circles) are repre-
sented vs. the pH. (C) The low-pH region of the steady-state fluores-
cence experiments (the symbols are the same as in A). The conditions
were: 6 l
M
of protein, and 40 l
M
ANS, when required, at
25 (± 0.1) °C; buffer concentration was 10 m
M
in all cases; spectra

were acquired in 0.5-cm pathlength cells. The lines are fittings to
Eqn (2).
2258 G. Ferna
´
ndez-Ballester et al.(Eur. J. Biochem. 270) Ó FEBS 2003
Conversely, as the pH increased, the maxima wavelength
and the spectral intensities remained constant up to high pH
(Fig. 1A).
Theprofileof<k> vs. pH (Fig. 1A) showed a sigmoidal
behaviour at low pH, and a plateau region above pH 4. The
maxima wavelength followed the same pattern than the
<k>. The apparent pK
a
was 3.2 ± 0.3 (Fig. 1A).
The protein without the His
6
-tag showed the same
behaviour at the different pH values than that observed
for the His
6
-tagged protein (data not shown).
Examination of tryptophan exposure by fluorescence
quenching. To further check whether the tertiary structure
around the sole Trp residue changes upon pH, we examined
iodide and acrylamide quenching by excitation at 280 nm,
which provided us with information about burial of the
indole moiety. Acrylamide-quenching experiments, carried
out at pH 3.3 and 7.5, yielded exponential Stern–Volmer
plots. At pH 7.5, where as judged by fluorescence, CD and
FTIR measurements the protein was folded, the K

sv
and v
were small, indicating burial of the tryptophan moiety
(Table 1). These results are in agreement with the observa-
tion that the maxima wavelength of spectra appeared at
337 nm at that pH. Conversely, at low pH, both quenching
parameters were larger, indicating solvent-exposure of the
aromatic ring. This is also in agreement with previous
observations on the shift of the maxima wavelength at low
pH values. Larger values of the quenching parameters were
also observed at pH 7.5 in the presence of 6
M
urea, where
the protein was completely unfolded. The fact that the
values of K
sv
in the presence and in the absence of urea, as
measured by acrylamide, are very similar within the error
(Table 1) is not fully understood, but it could be due to the
larger size of acrylamide when compared to that of KI.
Therefore, we can conclude that as the pH decreases, the
Trp is more solvent-exposed.
The use of KI as a quenching agent yielded similar results,
but experiments at very low pH values could not be carried
out because of precipitation. We do not know why HPr
precipitates, but it might be due to the presence of the
negative charge of the I
–
and the large amount of positive
charges at low pH, as happens in other molten-globule-like

species [34]. Here, conversely to that observed in acrylamide
experiments, only linear plots were found. A small K
sv
value
was found when the protein is folded, and larger values were
observed when the protein was completely unfolded (6
M
urea).
It is interesting to note here, that the K
sv
(in acrylamide
and KI) and v (in acrylamide) values for folded (and
unfolded) HPr are similar to those found in other folded
(an unfolded) proteins [35,36].
ANS binding fluorescence. ANS binding is used to
monitor the extent of exposure of protein hydrophobic
regions, and to detect the existence of non-native partially
folded conformations. When ANS is bound to solvent-
exposed hydrophobic patches of proteins, its quantum yield
is enhanced and the maxima of the emission spectra is
shifted from 520 nm to 480 nm [37]. At low pH values, the
intensity of the ANS in the presence of HPr was largely
enhanced and the maxima wavelength appeared at 482 nm
(Fig. 1B). As the pH was increased, the spectra intensity was
reduced and the maxima wavelength shifted towards
528 nm. These results suggest that: (a) ANS was bound to
HPr at low pH values, probably because of the presence of
solvent-exposed hydrophobic regions and (b) those hydro-
phobic patches were probably buried in the pH range 4–7,
as concluded from the titration curve measured (Fig. 1B).

The apparent pK
a
was 5.3 ± 0.5. The ANS-binding
experiments carried out with the protein without the
His
6
-tag showed the same behaviour (data not shown).
Far-UV and near-UV CD. We used far-UV CD in the
analysis of the unfolding of HPr as a spectroscopic probe
that is sensitive to the presence of protein secondary
structure [22]. Its CD spectrum was intense and showed
the typical features of an a-helical protein, with intense
minima at 222 and 208 nm (Fig. 2A), although interference
from the absorbance of tryptophan and histidine residues
could not be ruled out at 222 nm [22]. The shape of the CD
spectrum of S. coelicolor HPrwassimilartothatobserved
for E. coli HPr [16]. This shape did not change substantially
over the pH range explored (Fig. 2A), but the intensity at
low pH values was smaller than at neutral pH values.
However, the ellipticity at low pH values (pH 3.0) was not
as small as that observed at high urea concentrations, where
the protein was completely unfolded suggesting that the
protein has residual structure at low pH values. The
apparent pK
a
was 3.5 ± 0.3 (Fig. 2B), which is, within
the error, the same as that determined by fluorescence.
We used near-UV CD to detect possible changes in
the asymmetric environment of aromatic residues [22]. The
near-UV of HPr was very weak and no intense bands were

observed (Fig. 2C), probably due to the low content of
aromatic residues. The spectrum showed a small band at
292 nm and a shoulder at 285 nm corresponding to the
vibronic components of the
1
L
b
transition in tryptophan
residues [38]. Because of lack of intense distinctive features,
we did not use further the near-UV spectrum to map any
temperature, pH or chemical denaturant changes.
FTIR spectroscopy. FTIR is a powerful method for
investigation of protein secondary structure. The main
advantage in comparison with CD is that FTIR is more
sensitive to the presence of b structure or random-coil.
Measurements were only carried out at selected pH values
from pH 2.5–7.5, because of the large amounts of protein
Table 1. Quenching constants for HPr in acrylamide and KI. Errors are
data fit errors to Eqn (1) with (acrylamide) or without (KI) the
exponential factor. The K
sv
and v were obtained by fitting of fluores-
cence intensity at 335 nm vs. concentration of quenching agent (similar
values were obtained by fitting the intensities at 336, 337 and 338 nm,
data not shown). Experiments were carried out at 25 °C. The value
with a Ô–’ could not be measured due to HPr precipitation.
Conditions
Acrylamide KI
K
sv

(
M
)1
) v (
M
)1
) K
sv
(
M
)1
)
pH 3.3 4.5 ± 0.3 3.0 ± 0.1 –
pH 7.5 3.4 ± 0.3 2.0 ± 0.2 0.82 ± 0.05
6
M
urea 4.2 ± 0.5 7.9 ± 0.2 4.5 ± 0.1
Ó FEBS 2003 Stability of Streptomyces coelicolor HPr (Eur. J. Biochem. 270) 2259
used in the FTIR experiments and the impossibility of
sample recovery. As the pH increased the maxima of the
amide I bands moved from 1651.0 to 1645.6 cm
)1
(Fig. 3B).
The titration, as concluded from the position of the amide I
band, followed the same sigmoidal behaviour than that
observed by CD and fluorescence, but a reliable value for
the pK
a
midpoint could not be determined.
Deconvolution of the bands at the extreme explored pH

values (pH 2.5 and 7.5) indicates that the helical structure
remained basically unaltered (50%); conversely, antiparallel
b structures (the 1628 and the 1666 cm
)1
bands) were
increased at lower pH values (it changed from 20% at
pH 7.5–27% at low pH). Although the exact reasons of the
increase in the b sheet bands are unknown, it could be due to
formation of soluble oligomers at low pH values. The
presence of those oligomer species would explain also the
apparent thermal titration observed at low pH values in
the FTIR experiments.
Gel-filtration chromatography. Gel filtration provides a
measurement of the compactness (hydrodynamic volumes)
of the polypeptide chain [39]. HPr eluted at neutral pH at
Fig. 2. CD of HPr under different conditions. (A) Far-UV CD spectra at different conditions (filled circles, pH 3; open circles, neutral pH; squares,
6
M
urea). The conditions were: 20 l
M
of protein at 25 (± 0.1) °C; buffer concentration was 10 m
M
inallcases;spectrawereacquiredin0.1-cm
pathlength cells. (B) pH-induced unfolding of HPr following the ellipticity at 222 nm in the far-UV CD. The line is the fitting to Eqn (2). (C) Near-
UV of HPr. Continous line (circles) is the near-UV spectra at pH 7.5, 10 m
M
phosphate buffer. Dotted line (squares) is the near-UV at 6
M
urea,
pH 7.5, 10 m

M
phosphate buffer. Conditions were: 60 l
M
protein, 25 °C, spectra were acquired in 0.5-cm pathlength cells. (D) CD spectra at
different concentrations at pH 3.0, after normalization by the smaller concentration used (10 l
M
). Open circles (black lines) are spectra acquired at
10 l
M
; open squares (red lines) are at 20 l
M
and filled circles (blue lines) at 40 l
M
protein concentration. The ellipticity units on the y-axis are the
normalized raw ellipticity.
Fig. 3. FTIR of HPr at different pH values.
(A) Green (pH 2.5), blue (pH 3), red (pH 5)
and black lines (pH 7). Vertical bars indicate
absorbance units. (B) Position (cm
)1
)ofthe
amide I band at different pH values. Protein
concentration was 6 mgÆmL
)1
; all other con-
ditions as described under the Experimental
procedures section.
2260 G. Ferna
´
ndez-Ballester et al.(Eur. J. Biochem. 270) Ó FEBS 2003

the volume expected for a globular folded protein of its size,
13.92 mL. The R
s
determined from Eqn (6) is 16.35 A
˚
,
within the range expected for a globular protein of its size
(Fig. 4A). At low pH values, the protein eluted close to
the void volume of the column as a single peak (Fig. 4B),
8.55 mL at pH 3.5. From pH 4–4.5, HPr eluted in two
different peaks: (a) the first peak eluted at the same volume
as the pH-unfolded protein and (b) the second peak was
very broad. These data indicate that the interconversion
between the unfolded and native forms of the protein is slow
as compared to the column retention time [39]. Similar
behaviour has been observed in other partially folded
species [40].
The fact that the elution volume of HPr at low pH values
is close to the void volume suggests that either HPr is an
oligomer or it has lost most of this globular shape. We can
rule out the first explanation as: (a) glutaraldehyde cross-
linking at pH 3.0 did not show the presence of oligomers at
the concentrations explored, 20–100 l
M
(data not shown);
(b) far-UV experiments at different concentrations, ranging
from 14 to 295 l
M
, at pH 3.0 did not show any change
either in the shape or the raw ellipticity after being

normalized by the smallest concentration used (Fig. 2D);
(c) if there was a large population of aggregated forms, two
intense bands at 1620 cm
)1
and 1685 cm
)1
[30,31] should
appear in the FTIR spectrum; at pH 2.5, even at the large
amounts of protein used in the FTIR experiments, these
bands were not found at 25 °C (Fig. 6B); however, the
presence of an small amount of aggregated forms could
explain the increase (7%) in the b sheet structure observed
at this pH; (d) the gel filtration experiments carried out
at different protein concentrations (20–60 l
M
)atthese
low pH values did not show any difference in the elution
volume and (e) the <k> is very sensitive to changes in
oligomerization processes [23], and the value observed at
thesepHvaluesisclose(2.84lm
)1
) (Fig. 1A) to that
observed for urea-unfolded HPr (where aggregated forms
are not present); furthermore, only below pH 2 a large
increase in the <k> was observed (from pH 0.22–0.87,
<k> shifted from 2.90 to 2.86 lm
)1
, respectively); also, the
maxima wavelength decreased as the pH was reduced (from
pH 0.22–0.87, it shifted from 340 to 348 nm, respectively)

(Fig. 1C); these findings indicate the presence of aggregated
forms at pH values below 2. Then, all these probes suggest
that at pH 3.0, HPr is monomeric, and the small elution
volume indicates that HPr has lost most of its globular
form. Below this pH small populations of aggregated
species cannot be ruled out, but their contribution to
the spectral properties is insignificant as suggested by the
FTIR spectra.
Thermal-denaturation experiments at different
pH values
In order to obtain the thermodynamic parameters charac-
terizing the unfolding transition of HPr, we carried out
thermal-denaturation experiments followed by CD, FTIR
and DSC. Measurements trying to obtain a complete set of
thermodynamic parameters by using fluorescence failed due
to the large temperature dependence of the intrinsic
Fig. 4. pH-induced unfolding of HPr followed by gel-filtration chromatography. (A) Determination of the Stokes radius by gel filtration chroma-
tography on a HR Superdex G75 (Amersham Biosciences). The elution volume of HPr is indicated by an arrow and a square. The numeration
corresponds, respectively, to the elution volumes of ribonuclease A (1), chymotrypsinogen A (2), ovalbumin (3), and albumin (4). (B) Elution
volume vs. pH, the filled circles observed from pH 4–4.5 indicate one of the two peaks observed during protein elution. The open circles and squares
indicate the elution volumes of two different measurements. (C) Chromatograms at selected pH values: continuous black lines and open circles,
pH 3; dotted blue lines and open squares pH 4.3; dotted-and-dashed red lines and filled circles pH 7. Conditions were: 20–60 l
M
of protein at
25 °C, in 10 m
M
of the corresponding buffer and 150 m
M
NaCl.
Ó FEBS 2003 Stability of Streptomyces coelicolor HPr (Eur. J. Biochem. 270) 2261

fluorescence of both the native and unfolded states of the
protein (data not shown).
Far-UV CD. We explored the thermal-denaturation beha-
viour of HPr from pH 2.0–11.0 (Fig. 5), by following the
ellipticity at 222 nm. We found four different pH regions,
according to the reversibility and sigmoidal behaviour. (a)
Between pH 2.0 and 3.0 the heating transition did not have
a sigmoidal behaviour, but it was reversible. (b) Between
pH 3.5 and pH 4.5 the heating transition was noncooper-
ative and irreversible. (c) Between pH 5.0 and pH 5.5 the
heating transition showed a sigmoidal behaviour, but was
not reversible. (d) Between pH 6 and pH 9 the heating
transition had a sigmoidal behaviour and it was reversible.
The reversibility at those pH values was approximately
90–100%, as measured from the relative ellipticity recovery
in the reheating experiments. At higher pH values, the
thermal transitions were not reversible, probably due to
deamidation processes [16,41]. A reliable determination of
the DC
p
and DH
cal
between pH 6.0–9.0 could not be carried
out, because of the absence of baseline in the unfolded state
at some pH values. The thermal unfolding of HPr was
pH-independent from pH 7.0–9.0, as concluded from the
identical T
m
values (64.6 ± 0.6 °C), suggesting that protein
stability was similar in that pH range.

It is interesting to note here that the transition at pH 3.5
(Fig. 5A) showed a low degree of cooperativity at low
temperatures, although not a proper and complete sig-
moidal behaviour was observed. This could suggest the
presence of a molten-globule species, as has been seen in
a-lactalbumin and other proteins [42].
FTIR spectroscopy. Upon heating, the shape of the
amide I band of HPr changed dramatically at pH 2.5 and
7.5, as shown by: (a) a substantial loss in the integrated
intensity of bands arising from ahelix and (b) the
appearance of a strong and weak bands at 1620 and
1680 cm
)1
, respectively, which correspond to interactions
between extended chains, and have been related to aggre-
gation of thermally unfolded proteins [30] (Fig. 6). The
measurement of the whole band width at half-height upon
temperature allowed the characterization of the melting
curve at pH 7.5, which resembled that found by CD
experiments (Figs 5 and 6). Similar sigmoidal transitions
were also obtained by following the change in intensity in
the bands at 1652 cm
)1
(where the a helix appears) and at
1630 cm
)1
(where the b sheet is absorbing) (data not
shown). At pH 7.5, the transitions were irreversible,
precluding the determination of T
m

. The lack of reversibi-
lity, when compared to CD and DSC results, was probably
due to the high protein concentrations used.
At pH 2.5 a thermal transition was also observed
(Fig. 6A), in contrast to the experimental findings obtained
by CD. The presence of this transition at pH 2.5 is not
understood, but it could be due to aggregation processes
occurring at the high protein concentrations and tempera-
tures used in the FTIR experiments.
DSC experiments. We studied the heat-induced denatur-
ation of the protein by DSC at pH 7.5. The protein
(1 mgÆmL
)1
,87.5l
M
) was heated at a constant scanning
rate (60 °CÆh
)1
)upto95°C (scan), cooled down, and
reheated under identical conditions (re-scan). The scan and
re-scan experiments are equally well fitted by the two-state
model [43] with van’t Hoff to calorimetric enthalpy ratios,
DH
VH
/DH
cal
, of 1.02 and 1.01, respectively. The DSC
results indicate that the heat-induced unfolding of HPr was
characterized, under these experimental conditions, by a
melting temperature, T

m
, of 65.4 ± 0.5 °C, a calorimetric
enthalpy change upon unfolding, DH
cal
¼ 60.3 ± 1.5
kcalÆmol
)1
and an entropy change upon unfolding:
DS(T
m
) ¼ DH(T
m
)/T
m
¼ 177.8 calÆK
)1
Æmol
)1
(Fig. 7).
Chemical denaturation experiments
As the stability of HPr does not change significantly
around neutral pH, as concluded from thermal denatur-
ation experiments, we decided to follow the chemical
denaturation at pH 7.5 to compare with the stability
Fig. 5. Thermal denaturation profiles of HPr followed by far-UV CD at
222 nm. Continuous line (circles) is the heating experiment and the
dotted line (squares) is the reheating scan at (A) pH 3.5; (B) pH 5; and
(C) pH 7.5. The ellipticity units on the y-axis are arbitrary. The con-
ditions were: 20 l
M

of protein; buffer concentration was 10 m
M
;
spectra were acquired in 0.1-cm pathlength cells. The scan rate was
60 °CÆh
)1
in all cases.
2262 G. Ferna
´
ndez-Ballester et al.(Eur. J. Biochem. 270) Ó FEBS 2003
results obtained in other HPr family members. Figure 8
shows the chemical denaturation curves of HPr at pH 7.5,
10 m
M
phosphate, followed by fluorescence, far-UV and
gel filtration chromatography at 25 °C. The agreement
between the three probes suggests that the chemical-
denaturation can be described as a two-state model.
Thermodynamic parameters from gel-filtration and CD
measurements had a large error. We do not know the
reasons for those large errors, but they could be due to the
spread observed in the baselines of the transitions. Similar
large slopes in chemical denaturation experiments followed
by NMR have also been observed by following the
GdmCl chemical-denaturation in the E. coli HPr [16], but
they do not yield large errors in the thermodynamic
parameters. Chemical denaturation followed by ANS
binding did not show any sigmoidal titration (data not
shown), suggesting that no intermediate with close solvent-
exposed hydrophobic patches accumulated during the

denaturation. The free energy, determined from the
fluorescence measurements using the linear extrapolation
method approach, yielded a value of 4.0 ± 0.2 kcalÆmol
)1
,
indicating that HPr is not a highly stable protein. The
denaturation experiments were reversible in all cases.
Discussion
Equilibrium-unfolding of HPr at neutral and high
pH values follows a two-state mechanism
The chemical-denaturation folding of S. coelicolor follows a
two-state mechanism at neutral and high pH values, as
happens in other small proteins [44]. The denaturation of
HPr can be described as a two-state reaction at neutral pH
from the following evidence. (a) All the unfolding data can
be fitted to a single transition curve using Eqn (3). (b) The
denaturant transitions appear to be independent of the
biophysical probe (fluorescence, far-UV CD and gel-filtra-
tion chromatography) used (Fig. 8), and reversible for
either folding and unfolding (data not shown). (c) The ratio
of the van’t Hoff enthalpy of denaturation and the
calorimetric enthalpy obtained from DSC is close to unity
for the heating and the reheating scans (Fig. 7).
The conformational stability of S. coelicolor is the
smallest among that of the other family members reported
so far (Table 2). This low stability is also confirmed by
hydrogen exchange measurements at pH 7.5 using FTIR
and NMR experiments, where most of the protons
exchanged within 20 min (data not shown). Comparison
of the sequence among 31 HPr homologues in the protein

data bank indicates that the HPr of S. coelicolor forms a
completely different cluster of sequence equally distant to
the rest of the members of the family [9] (Fig. 9). The
differences in stability, as the structure of HPr of S. coeli-
color (unpublished results) is nearly identical to those of
E. coli and B. subtilis, must rely on differences in the
packing of their different side-chains.
Fig. 7. Excess heat capacity function of HPr at pH 7.5 in 10 m
M
phosphate buffer. The continuous lines represent the fitting of the
experimental data to a two-state reversible model.
Fig. 6. Thermal denaturation profiles of HPr followed by FTIR. (A)
Thermal denaturation profiles of HPr at pH 2.5 (filled squares) and
pH 7.5 (filled circles). The lines at both pH values are drawn to guide
the eye. (B) Thermal denaturation profiles of HPr at pH 7.5 (left side)
and pH 2.5 (right side) at selected temperatures. Protein concentration
was 6 mgÆmL
)1
; all other conditions as described under the Experi-
mental procedures section.
Ó FEBS 2003 Stability of Streptomyces coelicolor HPr (Eur. J. Biochem. 270) 2263
Unfolding of HPr at acid pH values
From pH 2.5–4.0, the behaviour of the different biophysical
probes suggests that HPr is sampling other conformations,
different to the folded and unfolded states. We have
monitored the pH unfolding of HPr by several biophysical
techniques and found that the curves cannot be super-
imposed. This is classical evidence of the accumulation of an
equilibrium intermediate [46]. It is worth mentioning here
that this is the first time such an intermediate has been

observed among the members of the HPr family.
At low pH values, the tryptophan is completely solvent-
exposed, as judged by the fluorescence and quenching
experiments (Table 1). It seems that the tertiary structure
around the Trp is lost. Conversely, there is some residual
secondary helical structure, as suggested by CD and FTIR
experiments. The residual ellipticity at 222 nm is )5200
degÆcm
2
Ædmol
)1
(pH 2.5) vs. the )6900 degÆcm
2
Ædmol
)1
(pH 7.5) (Fig. 2). However, the FTIR experiments indicate
that the amount of a-helical structure is nearly the same as
that in the native state (approximately 50% of the whole
percentage of structure). The difference between both
techniques must rely either on: (a) the FTIR deconvolution
procedures or (b) possible rearrangements of the aromatic
residues, which also absorb at 222 nm in the CD spectra.
Then, the probes indicate that at low pH values the tertiary
native-like structure is lost, but there is still some secondary
structure present. However, we do not know whether the
residual secondary structure is native or non-native.
As the pH increases, the protein goes from a state with
residual secondary structure, but no tertiary-like structure,
to a conformation with native-like secondary and tertiary
structure (Fig. 1). This acquisition of tertiary structure

follows a sigmoidal behaviour, as concluded from CD and
fluorescence spectra. The low pK
a
value of the transition,
when compared to those of Asp, Glu or C-terminal residues
in random-coil models [47], suggests that the acid unfolding
of HPr is linked to an acid(s) residue(s) involved in
hydrogen-bonding, salt-bridges or surrounded by a large
amount of positive charged residues.
At pH 4, the polypeptide chain has attained a native-like
secondary structure, and a native-like environment around
the indole moiety, as concluded by the native-like maxima
of fluorescence spectra and the native ellipticity (Figs 1 and
2). However, the compactness is not native-like, as judged
by the gel-filtration experiments. In a narrow pH range (i.e.
from pH 4–4.5) the protein attains its native compactness.
We do not know whether the acquisition of native-like
compactness of the polypeptide chain is a consequence of
Fig. 8. Urea denaturation curves of HPr at
pH 7.5, 25 °C. Changes are plotted for fluor-
escence (filled circles and black lines), CD
(open circles and blue lines), and gel filtration
chromatography (filled squares and red lines),
as the fraction of native protein. The thermo-
dynamic parameters were: fluorescence:
m (kcalÆmol
)1
Æ
M
)1

) ¼ 1.27 ± 0.04; [urea]
50%
¼
3.16 ± 0.02
M
;CD:m (kcalÆmol
)1
Æ
M
)1
) ¼
1.3 ± 0.1; [urea]
50%
¼ 2.97 ± 0.09
M
;and
gel filtration: m (kcalÆmol
)1
Æ
M
)1
) ¼ 1.5 ± 0.2;
[urea]
50%
¼ 2.9 ± 0.2
M
.
Table 2. Thermodynamic parameters of HPr family members.
Parameter E. coli HPr
a

B. subtilis
HPr
b
S. coelicolor
HPr
c
DH
m
(kcalÆmol
)1
) 70.8 ± 3.4 58.1 ± 1.4 60.3 ± 1.5
T
m
(°C) 63.8 ± 0.5 73.4 ± 0.2 65.4 ± 0.5
m (kcalÆmol
)1
Æ
M
)1
) 1.16 ± 0.11 1.08 ± 0.05 1.27 ± 0.04
[urea]
50%
4.58 ± 0.04 3.90 ± 0.05 3.16 ± 0.02
DG (kcalÆmol
)1
)
d
5.3 ± 0.5 4.2 ± 0.2 4.0 ± 0.2
a
Data are from Nicholson and Scholtz [45] at 10 m

M
phosphate
buffer, pH 7.0, 25 °C. There is another complete set of data for the
E. coli HPr, obtained from Dobson and coworkers [16], which are
similar to those described here, but obtained in the presence of
GdmCl. The value of the DG obtained from that set of data is
4.9 ± 0.4 kcalÆmol
)1
.
b
Data are from Scholtz [24] at 10 m
M
phosphate buffer, pH 7.0, at 25 °C.
c
Data from the fluorescence
and DSC experiments in this work; experiments were carried out at
10 m
M
phosphate buffer, pH 7.5, at 25 °C.
d
The free energy was
obtained from DG ¼ m [urea]
50%
.
Fig. 9. Sequence alignment. Aligned sequen-
ces of HPr of E. coli, B. subtilis and S. coeli-
color. The initial methionine in all sequences
has not been indicated.
2264 G. Ferna
´

ndez-Ballester et al.(Eur. J. Biochem. 270) Ó FEBS 2003
the previously completed titration (i.e. the native-like
compactness is not acquired until most of the molecules
have finished the low acid titration) or it is reflecting the
presence of other acidic groups, whose titration causes the
chain to acquire a native-like compactness. We favour,
however, the latter explanation as the thermal behaviour of
HPr between pH 3.5 and 4.5 is different to that observed at
lower pH values. Although it is difficult to rationalize a
native-like tertiary structure around the tryptophan without
the acquisition of a native-like compactness, it could be that
the sole tryptophan did not map correctly the acquisition of
tertiary structure, and further structural rearrangements
were required to attain a native-like tertiary compactness.
Preliminary NMR analysis at pH 4.5 indicate that the
protein has native-like chemical shifts for the tryptophan
moiety, but also non-native NOEs are observed (unpub-
lished results).
There is, still, a third titration towards the native state, as
shown by ANS binding (Fig. 1). Although the protein has
attained native-like compactness, secondary, and tertiary
structure, the ANS-binding experiments suggest that not all
the hydrophobic patches are buried. This titration could be
associated either with: (a) the own ANS titration or (b) the
presence of aggregated forms of the protein at those pH
values. Control experiments indicate that: (a) the fluores-
cence of ANS alone did not change significantly in the
whole pH range explored (data not shown); (b) when the
titration is occurring, the protein is not an oligomer, as
concluded from the gel filtration experiments (Fig. 4) and

(c) ANS-binding experiments at other protein concentra-
tions showed the same titration behaviour (data not shown).
Furthermore, pH-unfolding experiments in other proteins,
followed by ANS binding, have also shown titration
midpoints associated to the ionization of acid groups of
the macromolecule [48]. It seems, then, that the transition
mapped by ANS binding is indicating another partially
folded conformation, associated with either an acidic
residue or, the active-site His, whose pK
a
measured in other
HPrs, is small (pK
a
¼ 5.6) [49]. However, it could be
thought that the detected ANS transition in HPr was not
real because of the intrinsic spectroscopic and physical
properties of the probe itself. Two questions can be raised
concerning that point. Firstly, ANS has been shown to
induce partially folded species in some proteins [50].
Secondly, it has been suggested that there could be problems
in using emission maxima (or average energy) to track
pH-unfolding transitions, unless the fluorescence intensity
(quantum yield) of the two equilibrium states (e.g. the
well-folded and the low-pH protein forms) were approxi-
mately the same [51]. If one species (e.g. the low-pH form)
had a higher intensity than the other state, the transition
would be skewed, and the emission maximum would
overemphasize the more intensely emitting state. This
occurs because one of the states (e.g. the well-folded protein
state) binds ANS weakly and thus fluoresces weakly,

compared to the low-pH unfolded state. Then, the apparent
difference in the midpoint of the pH transitions (Fig. 1)
could be likely explained in this manner. However, both
objections can be ruled out considering that the thermal
denaturations of HPr in this pH range, despite the sigmoidal
behaviour during the heating scan, follow a different pattern
to that observed at lower and higher pH values. These
findings suggest that there are regions of the protein
which are not well-fixed and upon heating become more
disordered.
Then, the different biophysical probes suggest that at low
pH values, HPr acquires a partially folded conformation,
devoid of native-like tertiary structure, compactness and
with solvent-exposed hydrophobic patches, but with a large
amount of secondary structure. These intermediate species
would only be present at low pH values, as they are not
observed during the chemical-denaturation experiments at
neutral pH, as concluded from the evidence previously
discussed and the absence of ANS binding during urea
denaturation. However, we do not know if those species
play any role in the kinetic productive folding of HPr or are
simply off-pathway, dead-end, species. We can speculate
that probably those partially folded species might be present
in other HPr family members under other solution condi-
tions, but they would not have been observed previously
because of their larger conformational stability.
Is the partially folded conformation of HPr
a molten-globule-like structure?
In some cases, equilibrium intermediates characterized in
different proteins at low pH values have shown common

characteristics, such as: (a) the presence of a pronounced
amount of secondary structure; (b) the absence of most of
the native tertiary structure, as a result of lack of the tight
packing of side chains; (c) loosely packed hydrophobic core
that increases the hydrophobic surface accessible to solvent
and (d) almost a native-like compactness. This equilibrium
intermediate has been called Ômolten globule’, and it has
been proposed to be a general state of proteins. Further-
more, in some proteins, it resembles the kinetic intermedi-
ates found during the folding pathway [42].
The partially folded form observed in HPr at low pH
values (pH 2–4) has a large amount of secondary structure
(CD and FTIR experiments), the tertiary structure is lost (as
judged by fluorescence and quenching experiments, Fig. 1A
and Table 1), and it has solvent-exposed hydrophobic
regions (Fig. 1B). However, this partially folded form lacks
native-like compactness (as judged by the gel filtration
experiments, Fig. 4). Thus, the partially folded form of HPr
would not fit properly the definition of a classical molten
globule, because the absence of native-like compactness
suggests that its hydrophobic core is still very well-hydrated.
The second transition mapped by the ANS-binding pro-
perties does not fit either to a proper molten-globule
definition, because when HPr is still able to bind ANS, the
secondary and tertiary structure, and the compactness are
native-like. Nonetheless, it could be thought that because
there is only one fluorescence probe at the beginning of the
molecule, the acquisition of native-like tertiary structure by
fluorescence or quenching experiments were not well-
described. Then, in this scenario, ANS binding would

report the locking of the rest of nonfluorescent side-chains
and hydrophobic patches (i.e. the consolidation of the well-
fixed native tertiary structure). If this occurred, we could
suggest that from pH 4.5–6, HPr populates a molten-
globule state, with cooperative thermal unfolding denatu-
rations. This thermal cooperativity would indicate that
most of the native specific structure is present.
Ó FEBS 2003 Stability of Streptomyces coelicolor HPr (Eur. J. Biochem. 270) 2265
To sum up, we report in this work, for the first time, the
presence of partially folded species in the equilibrium
unfolding of one of the members of the HPr family. These
species are detected because of the smaller stability of
S. coelicolor HPr, when compared to other members of the
family.
Acknowledgements
We thank the two anonymous reviewers for helpful suggestions and
corrections. We thank Catalina Robledano and Sergio Gala
´
nfortheir
involvement in the early stages of this work. We thank F. N. Barrera
for helpful discussions and suggestions. We thank Xavier Avile
´
sfor
acquiring the mass spectra in his laboratory. We deeply thank May
Garcı
´
a, M. Carmen Fuster, Javier Casanova and M. T. Garzon for
excellent technical assistance. This work was supported by Projects
GV-00-024-5, and CTIDIB/2002/6 from Generalitat Valenciana;
Project BIO2000-1081, from the Spanish Ministry of Science; and

FIS01/0004-02 from the Spanish Ministry of Health.
References
1. Postma, P.W., Lengeler, J.W. & Jacobson, G.R. (1993) Phos-
phoenolpyruvate carbohydrate phosphotransferase system of
bacteria. Microbiol. Rev. 57, 543–594.
2. Bru
¨
ckner, R. & Titgemeyer, F. (2002) Carbon catabolite repression
in bacteria: choice of the carbon source and autoregulatory limi-
tation of sugar utilization. FEMS Microbiol. Lett. 209, 141–148.
3. Bentley, S.D., Chater, K.F., Cerdeno-Tarraga, A.M., Challis,
G.L., Thomson, N.R., James, K.D., Harris, D.E., Quail, M.A.,
Kieser, H., Harper, D. et al. (2002) Complete genome sequence of
the model actinomycete Streptomyces coelicolor A3 (2). Nature
417, 141–147.
4. Parche, S., Nothaft, H., Kamionka, A. & Titgemeyer, F. (2000)
Sugar uptake and utilisation in Streptomyces coelicolor:aPTS
view to the genome. Antonie Van Leeuwenhoek 78, 243–251.
5. Titgemeyer, F., Walkenhorst, J., Cui, X., Reizer, J. & Saier, M.H.
Jr (1994) Proteins of the phosphoenolpyruvate: sugar phospho-
transferase system in Streptomyces: possible involvement in the
regulation of antibiotic production. Res. Microbiol. 145, 89–92.
6. 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.
7. Butler, M.J., Deutscher, J., Postma, P.W., Wilson, T.J., Galinier,
A. & Bibb, M.J. (1999) Analysis of a ptsH homologue from
Streptomyces coelicolor A3 (2). FEMS Microbiol. Lett. 177,
279–288.

8. Kamionka, A., Parche, S., Nothaft, H., Siepelmeyer, J., Jahreis,
K. & Titgemeyer, F. (2002) The phosphotransferase system of
Streptomyces coelicolor: IIA-Crr exhibits properties that resemble
transport and inducer exclusion function of enzyme IIA-Glucose
of Escherichia coli. Eur. J. Biochem. 269, 2143–2150.
9. Parche,S.,Schmid,R.&Titgemeyer,F.(1999)ThePTSsystemof
Streptomyces coelicolor: identification and biochemical analysis of
a histidine phosphocarrier protein HPr encoded by the gene ptsH.
Eur. J. Biochem. 265, 308–317.
10. Hammen, P.K., Waygood, E.B. & Klevit, R.E. (1991) Reex-
amination of the secondary and tertiary structure of histidine-
containing protein from Escherichia coli by homonuclear and
heteronuclear NMR spectroscopy. Biochemistry 30, 11841–11850.
11. 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.
12. Maurer, T., Doker, R., Gorler, A., Hengstenberg, W. & Kalbitzer,
H.R. (2001) Three-dimensional structure of the histidine-con-
taining phosphocarrier protein (HPr) from Enterococcus faecalis
in solution. Eur. J. Biochem. 268, 635–644.
13. 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.NatlAcad.Sci.USA89, 2499–2503.
14. 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.
15. 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.
16. 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.
17. Gunasekaran, K., Eyles, S.J., Haggler, A.T. & Gierasch, L.M.
(2001) Keeping in the family: folding studies of related proteins.
Cur. Opin. Struct. Biol. 11, 83–93.
18. Pace, C.N. (1986) Determination and analysis of urea and
guanidine hydrochloride denaturation curves. Methods Enzymol.
131, 266–280.
19. Miroux, B. & Walker, J.E. (1996) Over-production of proteins in
Escherichia coli: mutant hosts that allow synthesis of some mem-
brane proteins and globular proteins at high levels. J. Mol. Biol.
260, 289–298.
20. 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.
21. Lakowicz, J.R. (1999) Principles of Fluorescence Spectroscopy,2nd
edn. Plenum Press, New York.
22. Woody, R.W. (1995) Circular dichroism. Methods Enzymol. 246,
34–71.
23. Royer, C.A. (1995) Fluorescence stability. In Protein Stability and

Folding (Shirley, B.A., ed), pp. 65–89. Humana Press, Towota,
New Jersey.
24. Scholtz, J.M. (1995) Conformational stability of HPr: the histi-
dine-containing phosphocarrier protein from Bacillus subtilis.
Protein Sci. 4, 35–43.
25. Schellman, J.A. (1987) The thermodynamic stability of proteins.
Annu. Rev. Biophys. Biophys. Chem. 16, 115–137.
26. Aune, K.C. & Tandford, C. (1969) Thermodynamics of the
denaturation of lysozyme by guanidine hydrochloride. II.
Dependence on denaturant concentration at 25 degrees. Bio-
chemistry 8, 4586–4590.
27. Hinkle,A.,Goranson,A.,Butters,C.A.&Tobacman,L.S.(1999)
Roles for the troponin tail domain in thin filament assembly and
regulation. A deletional study of cardiac troponin T. J. Biol.
Chem. 274, 7157–7164.
28. Ackers, G.K. (1967) Molecular sieve studies of interacting protein
systems. I. Equations for transport of associating systems. J. Biol.
Chem. 242, 3026–3034.
29. Darlin, P.J., Holt, J.M. & Ackers, G.K. (2000) Coupled energetics
of lambda cro repressor self-assembly and site-specific DNA
operator binding I: analysis of cro dimerization from nanomolar
to micromolar concentrations. Biochemistry 39, 11500–11507.
30. Fernandez-Ballester, G., Castresana, J., Arrondo, J.L.R., Ferra-
gut, J.A. & Gonzalez-Ros, J.M. (1992) Protein stability and
interaction of the nicotinic acetylcholine receptor with cholinergic
ligands studied by Fourier-transform infrared spectroscopy.
Biochem. J. 288, 421–426.
2266 G. Ferna
´
ndez-Ballester et al.(Eur. J. Biochem. 270) Ó FEBS 2003

31. Moffat, D.J. & Mantsch, H.H. (1992) Fourier resolution
enhancement of infrared spectra data. Methods Enzymol. 210,
192–201.
32. Plum, G.E. & Breslauer, K.L. (1995) Calorimetry of proteins and
nucleic acids. Cur. Opin. Struct. Biol. 5, 682–690.
33. Eftink, M.R. (2000) Use of fluorescence spectroscopy as a thermo-
dynamic tool. Methods Enzymol. 323, 459–473.
34. Fink, A.L. (1995) Compact intermediate states in protein folding.
Annu. Rev. Biophys. Biomol. Struct. 24, 495–522.
35. Dignam, J.D., Qu, X. & Cahire, J.B. (2001) Equilibrium unfolding
of Bombyx mori glycyl-tRNA synthetase. J. Biol. Chem. 276,
4028–4037.
36. Muzammil, S., Kumar, Y. & Tayyab, S. (2000) Anion-induced
stabilization of human serum albumin prevents the formation of
intermediate during urea denaturation. Proteins 40, 29–38.
37. Semisotnov, G.V., Rodionova, N.A., Razgulyaev, O.I., Uversky,
V.N., Gripas, A.F. & Gilmanshin, R.I. (1991) Study of the Ômolten
globule’ intermediate state in protein folding by a hydrophobic
fluorescent probe. Biopolymers 31, 119–128.
38. Gursky, O. (2001) Solution conformation of human apolipopro-
tein C-1 inferred from proline mutagenesis: far and near-UV CD
study. Biochemistry 40, 12178–12185.
39. Corbett, R.J.T. & Roche, R.S. (1984) Use of high speed size-
exclusion chromatography for the study of protein folding and
stability. Biochemistry 23, 1888–1894.
40. Neira, J.L. & Fersht, A.R. (1999) Acquisition of native–like
interactions in C-terminal fragments of barnase. J. Mol. Biol. 287,
421–432.
41. Sharma, S., Hammen, P.K., Anderson, J.W., Leung, A., Georges,
F.,Hengstenbeg,W.,Klevit,R.E.&Waygood,E.B.(1993)

Deamidation of HPr, a phosphocarrier protein of the phospho
enolpyruvate: sugar phosphotransferase system involves aspar-
agines 38 (HPr-1) and aparagine 12 (HPr-2) in isoaspartyl acid
formation. J. Biol. Chem. 268, 17695–11770.
42. Arai, M. & Kuwajima, K. (2000) Role of the molten globule state
in protein folding. Adv. Prot. Chem. 53, 209–282.
43. Martinez, J.C., Viguera, A.R., Berisio, R., Wilmanns, M., Mateo,
P.L., Filimonov, V.V. & Serrano, L. (1999) Thermodynamic
analysis of alfa-spectrin SH3 and two of its circular permutants
with different loop lengths: discerning the reasons for rapid folding
in proteins. Biochemistry 38, 549–559.
44. Jackson, S.E. (1998) How do small single-domain proteins fold?
Fold Des. 3, R81–R91.
45. 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 denatur-
ation. Biochemistry 35, 11369–11378.
46. Jamin, M. & Baldwin, R.L. (1996) Refolding and unfolding
kinetics of the equilibrium folding intermediate of apomyoglobin.
Nat. Struct. Biol. 3, 603–618.
47. Cantor, C.R. & Schimmel, P.R. (1980) Biophysical Chemistry.
W.H. Freeman Co, New York.
48. Genzor, C.G., Beldarraı
´
n, A., Go
´
mez-Moreno, C., Lo
´
pez-
Lacomba, J.L., Cortijo, M. & Sancho, J. (1996) Conformational

stability of apoflavodoxin. Protein Sci. 5, 1376–1388.
49. Anderson, W.J., Bhanot, P., Georges, F., Klevit, R.E. & Way-
good, E.B. (1991) Involvement of the carboxy-terminal residue in
the active site of the histidine-containing protein, HPr, of the
phosphoenolpyruvate: sugar phosphotransferase system. Bio-
chemistry 30, 9601–9607.
50. Kamen, D.E. & Woody, R.W. (2001) A partially folded inter-
mediate conformation is induced in pectate lyase C by the addition
of 8-anilino-1-naphtalenesulfonate (ANS). Protein Sci. 10,
2123–2130.
51. Eftink, M.R. (1994) The use of fluorescence methods to monitor
unfolding transitions in proteins. Biophys. J. 66, 482–501.
Ó FEBS 2003 Stability of Streptomyces coelicolor HPr (Eur. J. Biochem. 270) 2267