In the present work, and on the basis of pfLamA homology modeling prediction [2], residues Glu53 and Asp287 were replaced with alanine residues by site-directed mutagenesis, either indiv
Trang 1of Glu53 and Asp287 in Pyrococcus furiosus
endo-b-1,3-glucanase
Roberta Chiaraluce1, Rita Florio1, Sebastiana Angelaccio1, Giulio Gianese2,
Johan F T van Lieshout3, John van der Oost3and Valerio Consalvi1
1 Dipartimento di Scienze Biochimiche ‘A Rossi Fanelli’ Sapienza Universita` di Roma, Italy
2 Ylichron Srl c ⁄ o ENEA Casaccia Research Center, S Maria di Galeria, Italy
3 Laboratory of Microbiology, Wageningen University, the Netherlands
Endo-b-1,3-glucanase (EC 3.2.1.39) from the
hyper-thermophilic archaeon Pyrococcus furiosus (pfLamA) is
a laminarinase that displays considerable residual
ter-tiary structure in 7.9 m guanidinium chloride (GdmCl)
[1] A high DGH2 O value of 61.5 kJÆmol)1is associated
with the partial unfolding of pfLamA which, in
7.9 GdmCl, maintains the ability to bind calcium with
substantial recovery of native tertiary structure, a unique property of this enzyme [2]
pfLamA belongs to family 16 glycoside hydrolases [3], a family composed by 748 enzymes (http://www cazy.org/) According to their substrate specificity, the enzymes of this family can be assigned to different subgroups [4] (http://www.ghdb.uni-stuttgart.de⁄ )
Keywords
double mutant; glycoside hydrolases;
laminarinase; protein stability;
thermodynamic stability
Correspondence
V Consalvi, Dipartimento di Scienze
Biochimiche ‘A Rossi Fanelli’ Universita`
‘La Sapienza’, P.le A Moro 5, 00185 Rome,
Italy
Fax: +39 06 4440062
Tel: +39 06 49910939
E-mail: consalvi@caspur.it
(Received 26 July 2007, revised 8 October
2007, accepted 10 October 2007)
doi:10.1111/j.1742-4658.2007.06137.x
The thermodynamic stability of family 16 endo-b-1,3-glucanase (EC 3.2.1.39) from the hyperthermophilic archaeon Pyrococcus furiosus is decreased upon single (D287A, E53A) and double (E53A⁄ D287A) muta-tion of Asp287 and Glu53 In accordance with the homology model predic-tion, both carboxylic acids are involved in the composition of a calcium binding site, as shown by titration of the wild-type and the variant proteins with a chromophoric chelator The present study shows that, in P furiosus, endo-b-1,3-glucanase residues Glu53 and Asp287 also make up a calcium binding site in 7.9 m guanidinium chloride The persistence of tertiary structure in 7.9 m guanidinium chloride, a feature of the wild-type enzyme,
is observed also for the three variant proteins The DGH2 Ovalues relative to the guanidinium chloride-induced equilibrium unfolding of the three vari-ants are approximatelty 50% lower than that of the wild-type The destabilizing effect of the combined mutations of the double mutant is non-additive, with an energy of interaction of 24.2 kJÆmol)1, suggesting a communication between the two mutated residues The decrease in the thermodynamic stability of D287A, E53A and E53A⁄ D287A is contained almost exclusively in the m-values, a parameter which reflects the solvent-exposed surface area upon unfolding The decrease in m-value suggests that the substitution with alanine of two evenly charged repulsive side chains induces a stabilization of the non-native state in 7.9 m guanidinium chlo-ride comparable to that induced by the presence of calcium on the wild-type These results suggest that the stabilization of a compact non-native state may be a strategy for P furiosus endo-b-1,3-glucanase to thrive under adverse environmental conditions
Abbreviations
ANS, 8-anilinonaphthalene-1-sulfonate; BAPTA, 5,5¢-Br 2 -1,2-bis(O-aminophenoxy)ethan-N,N,N¢,N¢-tetraacetic acid; GdmCl, guanidinium chloride; pfLamA, endo-b-1,3-glucanase from Pyrococcus furiosus; SVD, singular value decomposition.
Trang 2which share the same b-jelly roll fold but display
nota-ble differences in their primary structure Twenty-three
crystal structures of members of this family have been
solved (http://www.cazy.org/); however, the crystal
structure of the laminarinase subfamily is still missing
[4,5] Seventeen out of the 23 available crystal
struc-tures demonstrate the presence of at least one metal
binding site (http://www.cazy.org/http://www.ghdb
uni-stuttgart.de⁄ ) [4,5] Sequence alignments of
pfLamA from different sources suggested that Asp287
[2,6], a residue conserved in most family 16 glycoside
hydrolases, may be part of a calcium binding site
(http://www.ghdb.uni-stuttgart.de⁄ ) According to this
hypothesis, a 3D homology model of pfLamA has
pre-dicted the presence of one or two potential binding
sites for metals and two pairs of negatively charged
amino acid residues have been assumed to be involved
in calcium binding: Glu53 and Asp287, and Glu239
and Glu246 [2]
In the present work, and on the basis of pfLamA
homology modeling prediction [2], residues Glu53 and
Asp287 were replaced with alanine residues by
site-directed mutagenesis, either individually (E53A,
D287A) or simultaneously (E53A⁄ D287A), in order to
demonstrate their involvement in calcium binding in
native conditions and in the presence of 7.9 m GdmCl
The thermodynamic stability of pfLamA variant
pro-teins has been studied by GdmCl-induced unfolding
equilibrium experiments The thermodynamic
charac-terization of the double mutant provided more
informa-tion than a study of single mutants, especially with
respect to the direct or indirect involvement of residues
Glu53 and Asp287 either in electrostatic interactions
with other protein residues or in metal binding [7]
Glu53 and Asp287 are negatively charged at neutral pH
and contribute to the optimization of electrostatic
charges balance of pfLamA in the native state,
indepen-dently of their interaction with calcium The role of
electrostatic interactions in protein stability has been
widely investigated and the stabilizing effect of salt
bridge networks on the native state of
hyperthermophil-ic proteins has been proposed on the basis of several
computational and experimental studies [8–12] Studies
of proteins from hyperthermophiles have provided an
array of hypotheses on the structural determinants
responsible for their resistance to denaturation [13,14];
however, a unifying description remains elusive [15]
Investigations on protein stability are also necessary to
advance our skills in designing new catalysts resistant to
temperature and extreme solvent conditions
In addition to their role in the stabilization of
pfLamA native state, Glu53 and Asp287 could also
contribute to the persistence of residual tertiary
struc-ture in the non-native state in 7.9 m GdmCl [1] The study of the properties of non-native states of proteins has received considerable attention in the last 10 years because the residual structure within the unfolded state may play an important role in a protein’s energetics and function [16–19] Changes in the denatured states induced by mutations can therefore affect protein sta-bility [20,21] and, in thermophilic proteins, the persis-tence of residual structure in non-native states may contribute toward avoiding irreversible denaturation under extreme environmental conditions [22] Under denaturing conditions, the spectroscopic characteriza-tion of the residual structure in proteins is very diffi-cult, although it has been reported in several studies [18,23–25]
A measure of the residual structure in the unfolded form of a protein is the thermodynamic parameter m,
as obtained in equilibrium unfolding studies [26–28] This parameter represents the rate of change of the free energy of unfolding as a function of denaturant concentration and is proportional to the amount of additional surface area exposed upon unfolding [26] The m-value may provide information about the resid-ual structure present in the denatured states of mutants
in comparison to that of the wild-type protein [26,29] Mutations affecting m-values are more likely to change the accessible surface area of the unfolded form rather than that of the native state; thus, a change in m-value
is generally considered to reflect a change in the com-pactness of the denatured state [26] The present study reports on the thermodynamic stability of pfLamA single mutants E53A, D287A and double mutant E53A⁄ D287A, as well as the binding of calcium to the mutant proteins in native conditions The single and combined mutations dramatically decrease the thermo-dynamic stability of the proteins with a significant decrease of m-values relative to the GdmCl-induced unfolding equilibrium A double-mutant thermody-namic cycle reveals a non-additive effect of the muta-tions on the thermodynamic parameters [30,31] The effect of the mutations indicates a key role for Glu53 and Asp287 in the interactions responsible for the residual structure in the non-native state, as well as for calcium binding of pfLamA in the native state and in 7.9 m GdmCl
Results
Spectroscopic characterization of the mutants in native conditions and in 7.9MGdmCl
In native conditions, the near-UV CD spectra of the three mutants E53A, D287A and E53A⁄ D287A are
Trang 3very similar to those of the wild-type enzyme, except
for minor differences in the ellipticity signals all
cen-tered around the same main aromatic bands of the
native wild-type (Fig 1A) The fluorescence emission
spectra of native wild-type and mutant enzymes are all
centered at the same maximum emission wavelength of
342 nm and have similar emission fluorescence
intensi-ties (Fig 1B) Analogously, the far-UV CD spectra are
virtually superimposable (data not shown) These
results indicate that the mutations had no effect on the
secondary and tertiary structure arrangements of the
protein and suggest that, in the native state, the effect
of the mutations are directed and localized to the
mutated residue
In the presence of 7.9 m GdmCl, and analogous to
that observed with the wild-type pfLamA [1,2], the
near-UV CD spectra of the three mutants indicate
the presence of substantial residual tertiary structure
(Fig 1C) Minor differences in the near-UV CD spec-tra of the mutants, in comparison to that of the wild-type, are evident in the 275–290 nm region, where the positive ellipticity is progressively reduced in the three mutants from the D287A to the double mutant, and in the 260–270 nm region where the negative ellipticity of the E53A⁄ D287A mutant is somewhat decreased (Fig 1C) The changes observed in the near-UV CD spectra of the three mutants indicate that, in 7.9 m GdmCl, the environment of the aromatic residues is slightly perturbed In particular, for the double mutant, the decrease in the dichroic activity around
260 nm suggests that Phe residues are locked differ-ently in their tertiary contacts compared to the wild-type (Fig 1C)
In 7.9 m GdmCl, the maximum fluorescence emis-sion wavelength of the three mutants is shifted to
357 nm, similar to the wild-type in the same
condi-Fig 1 Spectral properties of pfLamA wild-type and mutants: effect of GdmCl in the presence and absence of CaCl 2 Near-UV CD (A, C, E) and fluorescence (B, D, F) spectra of pfLamA wild-type (– Æ Æ –), D287A (—–), E53A ( ) and E53A ⁄ D287A (– ) –) were recorded at 20 C after 20 h of incubation of the protein in native conditions (20 m M Tris ⁄ HCl, pH 7.4) (A, B) and in 7.9 M GdmCl, pH 7.4, in the absence (C, D) and presence (E, F) of 40 m M CaCl 2 The spectral properties of all the proteins under native conditions are unchanged upon addition of
40 m M CaCl2(data not shown) Near-UV CD spectra (A, C, E) were recorded in a 1-cm quartz cuvette at 0.6 mgÆmL)1protein concentration Fluorescence spectra (B, D, F) were recorded at 40 lgÆmL)1protein concentration (290 nm excitation wavelength).
Trang 4tions, but the relative fluorescence intensities are
increased to a different extent (Fig 1D) The increase
in relative fluorescence intensity emission at 342 nm is
approximately two-fold for the wild-type and 2.5, 2.6
and 2.8-fold for the D287A, E53A and the double
mutant, respectively (Fig 1D) Noteworthy, similar to
that reported for the wild-type enzyme [1,2], the
far-UV CD spectra of the three mutants are not affected
by equilibrium incubation at increasing concentrations
of GdmCl up to 7.9 m (data not shown)
Equilibrium transition studies in GdmCl
The effect of increasing GdmCl concentrations (0–8 m)
on the structure of the three mutants was analyzed in
comparison to the effect exerted on the wild-type
pfLamA in 20 mm Tris⁄ HCl, pH 7.4, containing
100 lm dithiothreitol and 100 lm EDTA The intrinsic
fluorescence emission intensity of the three mutants
increases after 20 h of incubation at increasing GdmCl
concentrations (Fig 2) and, in 7.9 m GdmCl, the
max-imal fluorescence emission wavelength shifts to 357 nm
(Fig 1D) The changes in relative intrinsic fluorescence
emission intensity of the mutants show a sigmoidal
dependence on GdmCl concentration and follow a two-state denaturation process without any detectable intermediate, similar to that reported for the wild-type pfLamA (Fig 2) [2] The changes are fully reversible upon dilution of the denaturant, and the transition midpoints are at 6.2 ± 0.15 m for D287A and the double mutant, and at 6.0 ± 0.12 m for E53A (Fig 2), with the values being slightly lower than that
of the wild-type pfLamA, which is at 6.7 ± 0.13 m GdmCl (Fig 2) [2] Table 1 shows the thermodynamic parameters values obtained for wild-type and mutant forms of pfLamA The mg value of 9.2 kJÆmol)1ÆM)1
of the wild-type is approximately 30% lower than the value predicted from the number of the aminoacid res-idues (approximately 13 kJÆmol)1ÆM)1 for 263 amino acid residues) [28], in accordance with the persistence
of residual structure in 7.9 m GdmCl The mutants are thermodynamically less stable than the wild-type, with
a significant decrease of DGH 2 Oand mgvalues, suggest-ing that the mutations considerably affect the stability
of pfLamA The double mutant shows a slightly smal-ler stability than either the single mutants and the similarity between the DDGH2 O values of the variant proteins (Table 1) indicates that the double mutant E53A⁄ D287A is more stable than expected from the sum of the stability change from single mutants E53A and D287A, and hence the effect of the mutations is non-additive Calculation of the energy of interaction between two mutated residues, DDGint, according to Eqn (5), yields a value of 24.2 ± 1.87 kJÆmol)1
Effect of calcium on pfLamA mutants in 7.9M GdmCl and in native conditions
The addition of 40 mm CaCl2 to calcium-depleted samples of D287A and E53A in 7.9 m GdmCl causes changes in their tertiary structure, but these are much
Fig 2 GdmCl-induced fluorescence changes of pfLamA wild-type
(r, e), D287A (m, n), E53A (d, s) and E53A ⁄ D287A (j, h)
Con-tinuous lines are the nonlinear regression to Eqn (3) of the
fluores-cence data at varying denaturant concentrations, as described in
the Experimental Procedures The reversibility points (empty
sym-bols) were not included in the nonlinear regression analysis All
spectra were recorded at 20 C after 20 h of incubation at the
indi-cated GdmCl concentrations.
Table 1 Thermodynamic parameters for GdmCl-induced unfolding equilibrium of pfLamA wild-type and mutants All data were obtained at 20 C in 20 m M Tris ⁄ HCl, pH 7.4, containing 100 l M
dithiothreitol and 100 l M EDTA DG H 2 O and mg values were obtained from Eqn (3); [GdmCl] 0.5 was calculated from Eqn (4) Data are reported as the mean ± SE of the fit DDG H 2 O ¼ DG H 2 O
mutant ) DG H 2 O wild-type The SE value relative to DDG H 2 O was calculated according to: [SE(DDGH2 O)
]2¼ [SE(DG H2O
wild-type)]2+ [SE(DG H 2 O mutant)] 2
Protein
[GdmCl]0.5 ( M )
DGH2 O
(kJÆmol)1)
mg (kJÆmol)1Æ M )1)
DDGH2 O
(kJÆmol)1) Wild-type 6.7 ± 0.13 61.5 ± 1.23 9.2 ± 0.18 0 D287A 6.2 ± 0.15 36.5 ± 0.91 5.9 ± 0.15 )25.0 ± 1.53 E53A 6.0 ± 0.12 33.9 ± 0.68 5.6 ± 0.11 )27.6 ± 1.40 E53A ⁄ D287A 6.2 ± 0.15 33.1 ± 0.83 5.3 ± 0.13 )28.4 ± 1.48
Trang 5less pronounced than those observed for the wild-type,
indicating the involvement of Glu53 and Asp287 in the
interaction with the cation (Fig 1E,F) The regain of
aromatic chirality at 295 nm and in the 260–270 nm
region for D287A is similar to that observed for the
wild-type enzyme, whereas it is much less evident for
E53A (Fig 1E) With the double mutant, the near-UV
CD spectrum in 7.9 m GdmCl shows very minor
changes upon addition of CaCl2 (Fig 1E) The
intrin-sic fluorescence emission spectra of the mutants in
7.9 m GdmCl are affected by the presence of 40 mm
CaCl2 to different extents (Fig 1F) With D287A, the
intrinsic fluorescence emission intensity at 342 nm is
1.6-fold decreased, similar to the wild-type enzyme,
and the maximum emission wavelength is shifted to
347 nm, 4 nm higher than with the wild-type [2]
(Fig 1F) For E53A and the double mutant, the
changes of intrinsic emission fluorescence are less
evi-dent: the relative intensities are decreased 1.3-fold and
1.2-fold and the maximum emission wavelengths are
shifted to 350 nm with E53A and to 356 nm with the
double mutant, 7 nm and 13 nm more red-shifted than
that observed with the wild-type [2] (Fig 1F) The
far-UV CD spectra of the three mutants in 7.9 m GdmCl,
which are the same as those measured in the absence
of denaturant, are not affected by the addition of
CaCl2(data not shown)
Changes in both near-UV CD and fluorescence
spec-tra occur by tispec-tration in 7.9 m GdmCl with CaCl2,
from 0.2 nm to 150 mm unchelated Ca2+, although
with remarkably different amplitudes, depending on
the protein form (Fig 3) The amplitude of the
elliptic-ity changes at 295 nm decreases from the wild-type to
D287A and E53A and, above 200 lm Ca2+, no further
changes are observed (Fig 3A) For the double
mutant, only minor changes in the dichroic activity at
295 nm are detected over the whole range of the cation
concentration (Fig 3A) Nonlinear regression analysis
of the [Q]295 data for the wild-type pfLamA was used
to define two limiting slopes, intersecting at a value
which suggests that 2 mol of Ca2+per mol of enzyme
are required to reach an apparent saturation effect
(Fig 3A) [2] The changes in the fluorescence
proper-ties in 7.9 m GdmCl induced by CaCl2, represented by
a blue-shift of the maximum emission wavelength and
by a quenching of the fluorescence intensities (Fig 1F),
are reported in Fig 3B as the intensity-averaged
emis-sion wavelength (k) calculated according to Eqn (1)
and follow a hyperbolic dependence on CaCl2, similar
to that observed for [Q]295
In native conditions, the addition of 40 mm CaCl2
to the mutants did not affect the near-UV, the far-UV
CD and the fluorescence properties (data not shown),
similar to that reported for the pfLamA wild-type [2] The interaction of the mutants with calcium, in native conditions, was studied by titration with CaCl2 in the
Fig 3 Interaction of calcium with pfLamA wild-type and mutant forms in 7.9 M GdmCl (A) Left axis: [Q] 295 of wild-type (e) and D287A (n); right axis: [Q] 295 of E53A (s) and E53A ⁄ D287A (h) measured from near UV-CD spectra at 22 l M protein concentration (B) Intensity-averaged emission wavelength k of wild-type (e), D287A (n), E53A (s) and E53A ⁄ D287A (h) measured at 1.2 l M
protein concentration from fluorescence spectra (290 nm excitation wavelength) All spectra were recorded at 20 C, 5 min after each
Ca 2+ addition [Q] 295 is reported after removal of the high-frequency noise and the low-frequency random error by the singular value decomposition algorithm (SVD) [2] in the spectral region 250–
310 nm k was calculated according to Eqn (1) The two limiting slopes, calculated by nonlinear regression analysis to the [Q] 295 and
to k data, intersect at a point corresponding to [Ca 2+ unchelat-ed] ⁄ [protein] ¼ 2 The reported unchelated Ca 2+ concentrations intervals, calculated according to [49], are 0.2 n M to 140 m M and 0.2 n M to 7.6 m M for [Q] 295 and fluorescence changes, respec-tively.
Trang 6presence of the chromophoric chelator 5,5¢-Br2
-1,2-bis(O-aminophenoxy)ethan-N,N,N¢,N¢-tetraacetic acid
(BAPTA) and compared with the results obtained with
the wild-type enzyme [2] The fitting of the titration
data by caligator software [32] allows a quantitative
determination of the corresponding binding constants,
as reported in Table 2 All the mutants bind calcium
with a significantly lower affinity compared to the
wild-type Similar to that observed for the wild-type,
the v2 obtained by fitting the data of E53A and
D287A titration to a two calcium binding sites model
are both lower than that obtained by fitting the same
data to a one calcium binding site model (Table 2) In
the case of the double mutant, the v2value relative to
the fitting to a one calcium binding site model is lower
than that relative to the fitting to a two calcium
bind-ing sites model Furthermore, the higher value of the
second binding constant suggests that only one of the
two binding sites may be functional in E53A⁄ D287A
(Table 2)
Effect of calcium on the equilibrium transitions
in GdmCl
Incubation of pfLamA mutants at increasing GdmCl
concentrations (0–8 m) in 20 mm Tris⁄ HCl, pH 7.4,
containing 100 lm dithiothreitol, 100 lm EDTA and
40 mm CaCl2 for 20 h at 20C results in changes in
the intrinsic fluorescence emission (Fig 4) The
GdmCl-induced unfolding process in the presence of
40 mm CaCl2is reversible and, by contrast to the
wild-type (Fig 4A), does not follow a simple two-state
mechanism, as suggested by the lack of coincidence of
the changes in relative fluorescence intensity and in k
and by the hysteresis of the reversibility process
(Fig 4) At the end of the transition, the intrinsic
fluo-rescence emission intensity at 342 nm is increased
1.7-fold for D287A, 2.1-fold for E53A and 2.4-fold for
the double mutant and the fluorescence maximum
emission wavelength is shifted to 347 nm for D287A,
350 nm for E53A and to 356 nm for the double
mutant (Fig 4, insets) Notably, in 7.9 m GdmCl and
40 mm CaCl2, the maximum fluorescence emission wavelength of the wild-type was still centred at 342 nm [2] The fluorescence emission spectra of the three vari-ants measured after incubation in 7.9 m GdmCl and
40 mm CaCl2are comparable with those resulting from the progressive addition of CaCl2 to the proteins in 7.9 m GdmCl (Fig 1F, Fig 4, insets)
8-Anilinonaphthalene-1-sulfonic acid ammonium salt (ANS) fluorescence and acrylamide
quenching The amphiphilic dye ANS has affinity for hydrophobic clusters present in tertiary structure elements, which are not tightly packed within a fully folded structure The accessibility of hydrophobic residues of pfLamA wild-type and variant proteins in 7.9 m GdmCl was compared with that in native conditions, by analysis with the fluorescent probe ANS The fluorescence emission spectrum of ANS in the presence of all the variant proteins in 7.9 m GdmCl shows a modest, two-fold increase in intensity compared to that in native conditions, without any change in the maximum fluo-rescence emission wavelength (results not shown) This suggests that, in 7.9 m GdmCl, the hydrophobic sur-face area of the mutants is not significantly exposed, similar to that observed for the wild-type The uncharged fluorescence quencher acrylamide was used
to probe the accessibility of the hydrophobic core and the dynamic properties of the three mutants in com-parison to the wild-type in native conditions and in 7.9 m GdmCl Effective acrylamide quenching con-stants from the modified Stern–Vollmer plots for the proteins in the native state were 7.9 m)1, 8.6 m)1, 8.4 m)1 and 9.4 m)1 for the wild-type and D287A, E53A and the double mutant, respectively In 7.9 m
were 13.0 m)1, 9.9 m)1, 11.5 m)1 and 10.3 m)1 for the wild-type, D287A, E53A and the double mutant, respectively A quantitative analysis of the data is not
Table 2 Calcium binding constants for pfLamA wild-type and mutants determined in the presence of the chromophoric chelator BAPTA.
v2represents the best fit of the absorbance data Replicate determinations indicate a standard deviation for the calcium binding constants
K1and K2less than 5%.
Protein
2 ( M )1) v2
K 1 ( M )1) v2
Trang 7possible because pfLamA wild-type and variant
pro-teins are heterogeneously emitting systems; however,
the results indicate that the fluorophores accessibility
of the protein variants to the uncharged quencher, in
comparison with the wild-type, decreases in 7.9 m
GdmCl and increases in native conditions
Discussion
The results obtained with pfLamA mutant forms
indi-cate that residues Glu53 and Asp287 are involved in
calcium binding, in accordance to the homology
mod-elling The pfLamA single (D287A and E53A) and
double (E53A⁄ D287A) mutants in 7.9 m GdmCl show
a residual tertiary structure comparable to that of the
wild-type; however, the integrity of the calcium
bind-ing site formed by Asp287 and Glu53 is essential for
interaction with the cation in 7.9 m GdmCl
An interesting finding of the present study is the
sig-nificant decrease in the thermodynamic stability of the
three pfLamA mutants in comparison to the wild-type,
which shows a high DGH2 Ovalue of 61.5 kJÆmol)1
asso-ciated with its partial unfolding [2] The DGH2 O
associ-ated with the reversible fluorescence changes at
increasing GdmCl concentration is decreased from 1.7-,
1.8- to 1.9-fold with respect to the wild-type pfLamA,
going from D287A, to E53A and to the double mutant,
respectively Notably, the transition midpoints for the
fluorescence changes of the three mutants are not
sig-nificantly changed with respect to pfLamA wild-type;
hence, the decrease of DGH2 Ovalues is mainly due to a
decrease in mg values The decrease in mg value
observed in all the pfLamA variant proteins is
signifi-cant (1.6-fold) and not unprecedented for other single
[22,33,34] and double-mutant proteins [7] The
mecha-nism responsible for m– mutant proteins, which display
a mg value lower than that of the wild-type, is usually referred to a decrease in the solvent-exposed surface area upon unfolding This is more frequently ascribed
to an increase in the compactness of the residual struc-ture in the non-native state ensemble, rather than to an
Fig 4 GdmCl-induced fluorescence changes of pfLamA mutant
forms in the presence of CaCl 2 (A) D287A, (B) E53A and (C)
E53A ⁄ D287A Fluorescence changes are reported as relative
fluores-cence intensity at 342 nm (left axis: j, h) and as intensity-averaged
emission wavelength k (right axis: d, s) calculated according to
Eqn (1) The wild-type reversible transition in the presence of 40 m M
CaCl2monitored by relative fluorescence intensity at 342 nm (left
axis: e, r) is also shown in (A) for comparison [2] The solid lines
through the mutants unfolding data points (filled symbols) are
intended to guide the eye of the reader and do not represent the
fit-ting of the data Reversibility points are indicated by empty symbols.
All spectra were recorded at 20 C after 20 h of incubation at the
indicated GdmCl concentrations at 40 lgÆmL)1protein concentration.
Insets show intrinsic fluorescence emission spectra of the mutants
measured after 20 h of incubation in 7.9 M GdmCl and 40 m M CaCl 2
(continuous), the spectra resulting from the addition of CaCl2to the
mutants after 20 h of incubation in 7.9 M GdmCl (dotted) and the
spectra of the native mutants in 20 m M Tris ⁄ HCl pH 7.4 (dashed) All
spectra were recorded at 20 C (290 nm excitation wavelength).
Trang 8increase of the accessible surface area of the native state
[26,27,29] Similar to that reported for most m–mutants
[16], the spectral properties of the pfLamA variants in
7.9 m GdmCl do not indicate to a significant increase
in the structure of the non-native state to support the
significant decrease in mg Consistent with these results
and similar to that observed for the wild-type, the ANS
binding experiments indicate that, in 7.9 m GdmCl, the
hydrophobic surface area of pfLamA mutants is not
significantly exposed An increase in compactness of
the non-native state ensemble of the variants is
sug-gested by the decreased fluorophores accessibility to the
uncharged quencher acrylamide in 7.9 m GdmCl
com-pared to that of the wild-type In native conditions, the
spectral properties of the three variants point to tertiary
structures almost identical to that of the wild-type;
however, the higher fluorophores accessibility to the
uncharged quencher suggests a less compact native
state for the three mutant proteins A decrease in mg
value upon single mutation has been also referred, in
some cases, to the population of a third intermediate
state during chemical unfolding [35]; however, in our
experimental conditions, the presence of an
intermedi-ate stintermedi-ate was not observed for any of the pfLamA
vari-ants Thus, the decrease of mgvalue may be ascribed to
an increase in the compactness of the non-native state
in 7.9 m GdmCl
The effect of the double mutation on the
thermody-namic parameters was non-additive, being lower than
the sum of the effects of the two single mutations
Non-additivity is generally observed when two mutated
residues communicate directly or indirectly through
electrostatic interactions or structural perturbation, so
that they do not behave independently [7,36]
There-fore, the comparison of the thermodymamic
parame-ters of the pfLamA double mutant with those of single
mutants may provide information about any direct or
indirect interconnection between the two mutated
resi-dues The non-additivity of the stability change can be
expressed by the free energy coupling DDGint, a
param-eter calculated from a double-mutant cycle (Scheme 1)
that reflects the interaction energy between the two
mutated residues, Asp287 and Glu53 The positive
value of the interaction free energy between the two
mutated residues in pfLamA indicates that the double
mutant is more stable than predicted, on the
assump-tion that the effects of the two single mutants would
be additive A significant DDGint (above 20 kJÆmol)1)
has been related either to a direct communication
between two residues or a short range steric interaction
involving a mediating residue or a ligand [37] The
large interaction free energy between the two pfLamA
mutated residues (DDGint¼ 24.2 ± 1.87 kJÆmol)1) is
in accordance with the calcium binding results described in the present study The analysis of electro-static interactions in pfLamA model, including histi-dine residues and considering a distance threshold of
6 A˚, reveals that Glu53 may be involved in a large salt bridge network of seven ion pairs, three of which are strong ion pairs (distance threshold of 4 A˚) whereas the Asp287 is involved in three ion pairs, only one of which is a strong ion pair (Fig 5) The putative cal-cium binding site is localized in proximity of the well conserved Asp287 residue in family 16 [2,6] The homology model suggests ionic interactions between
Fig 5 Model structure of the calcium binding site of wild-type pfLamA The pfLamA model is represented in teal blue cartoons The two acidic residues binding calcium and those with which they form ion pair interactions are shown as stick models with superim-posed violet CPK space-filling models Carbon, oxygen and nitrogen atoms are displayed with green, red and blue colors, respectively Calcium ion is represented as an Nb sphere model with yellow color and a superimposed violet CPK space-filling model Orange dashes indicate ion pair interactions This figure was rendered using
PYMOL [50].
Scheme 1
Trang 9the cation and the carboxylate moiety of both Glu53
and Asp287 (Fig 5) In native conditions, the
spectro-scopic studies of pfLamA variant proteins with the
chromophoric chelator BAPTA show the involvement
of Glu53 and Asp287 in the interaction with calcium
and indicate that a second binding site might be
pres-ent, as predicted by the homology model [2] In 7.9 m
GdmCl, the capability to interact with calcium with
a consistent recovery of tertiary structure is still
observed, to a lesser extent than the wild-type, for
D287A and E53A, but not for D287A⁄ E53A Hence,
the integrity of the calcium binding site formed by
Asp287 and Glu53 is essential for interaction with
Ca2+ in 7.9 m GdmCl The second calcium binding
site revealed in native conditions by the chromophoric
chelator BAPTA and located by modeling between
Glu239 and Glu246 [2], either does not bind calcium
in 7.9 m GdmCl or it interacts with the cation without
affecting the enzyme tertiary structure
In the presence of calcium, the GdmCl equilibrium
transitions are complex and hysteretic for all the three
mutants, indicating that the cation may stabilize some
refolding intermediate(s) and⁄ or increase the
popula-tion of some states that are not evident under calcium
depletion The comparison with the simple two state
GdmCl transition of the wild-type in the presence of
calcium [2] suggests that the integrity of the cation
binding site formed by Glu53 and Asp287 prevents the
population of folding intermediates
The single or double substitution of Glu53 and
Asp287 by alanine decreases the capability of pfLamA
to interact with calcium as well as its thermodynamic
stability but not its intrinsic resistance to denaturation,
as indicated by the minor differences in the transition
midpoints The destabilizing effect appears to be
mainly realized through a stabilization of the
non-native state in 7.9 m GdmCl, rather than a
destabiliza-tion of the native state, as suggested by the decrease
of mg and of fluorophores accessibility to acrylamide
for all the variants compared to the wild-type The
replacement by Ala of any of the negatively charged
residues shown to be involved in the composition of a
calcium binding site induces a stabilization of the
non-native state in 7.9 m GdmCl comparable, but not
iden-tical, to that exerted by calcium on the wild-type [2]
Ion pair networks play an important role in protein
stability [8], and their involvement in interactions with
cations offers new perspective in the analysis of the
contribution of ions as cofactors in protein folding [38]
and in the design of variants of proteins with enhanced
stability [14] Changes in the denatured states induced
by mutation affect protein stability [20,21] and, for
thermophilic proteins, the persistence of residual
structure in non-native states may contribute to avoid irreversible denaturation under extreme environmental conditions [22] The stabilization of a compact non-native state may represent a strategy for P furiosus endo-b-1,3-glucanase to thrive under the most adverse environmental conditions
Experimental procedures
Site-directed mutagenesis
The pfLamA D287A mutant was prepared by overlap exten-sion PCR [39], using the wild-type construct pET9d::LamA
TTGC-3¢ (antisense) The E53A mutant and the double
primers 5¢-GCACGATGCGTTTGAAGG-3¢, and its com-plementary oligonucleotide The mutated bases are
construct pET9d::LamA [6] and the mutant construct
(La Jolla, CA, USA) The kit employs double-stranded DNA as template, two complementary oligonucleotide prim-ers containing the desired mutation, and DpnI endonuclease
to digest the parental DNA template Oligonucleotides were synthesized by MWG-Biotech AG (Anzinger, Germany)
The coding regions of the mutated pfLamA gene were sequenced to confirm the mutations and then E coli strain HMS174 (DE3) cells were transformed and used for expression
Enzyme preparation and assay
BL21(DE3) strain, and the three mutant forms were func-tionally produced in E coli HMS174(DE3) strain and puri-fied according to Kaper et al [40] The conditions used for expression and purification of the mutant proteins in E coli were as described for the wild-type enzyme The protein concentration was determined for wild-type and mutant
according to Gill and von Hippel [41] The enzyme activity was determined by measuring the amount of reducing sug-ars released upon incubation in 0.1 m sodium phosphate
Calcium-depleted protein was obtained by extensive dialysis with
Trang 10contamination were followed during the preparation and
storage of protein and buffer solutions [32] Calcium-loaded
Chemicals and buffers
ANS, dithiothreitol, EDTA, GdmCl, and laminarin were
from Fluka (Buchs, Switzerland) 3¢,5¢-dinitrosalicylic acid
was purchased from Sigma (St Louis, MO, USA) BAPTA
was from Molecular Probes Europe BV (Leiden, the
Neth-erlands) Buffer solutions were filtered (0.22 lm) and
care-fully degassed All buffers and solutions were prepared
with ultra-high quality water (ELGA UHQ, High
Wy-combe, UK) Buffers for calcium titrations were prepared
as previously described [32]
Spectroscopic techniques
Intrinsic fluorescence emission and 90 light scattering
mea-surements were carried out with a LS50B Perkin Elmer
spectrofluorimeter (Perkin Elmer, Waltham, MA, USA)
using a 1.0-cm pathlength quartz cuvette Fluorescence
emission spectra were recorded at 300–450 nm (1 nm
at 290 nm 90 light scattering was measured at 20 C with
both excitation and emission wavelength set at 480 nm to
check for the presence of aggregated particles
Far-UV (185–250 nm) and near-UV (250–320 nm) CD
1.0-cm pathlength quartz cuvette, respectively CD spectra
were recorded on a Jasco J-720 spectropolarimeter (Jasco
Inc., Easton, MD, USA) The results are expressed as the
mean residue ellipticity [Q] assuming a mean residue weight
of 110 per amino acid residue In all the spectroscopic
mea-surements, 100–250 lm EDTA was always present unless
otherwise stated
Experiments with the fluorescent dye ANS were
wild-type and variant proteins, with ANS at 1 : 20 molar ratio
After 5 min, fluorescence emission spectra were recorded at
400–600 nm with the excitation wavelength set at 390 nm
the intensity of the hydrophobic probe ANS depend on the
environmental polarity (e.g on the hydrophobicity of the
accessible surface of the protein) [42]
Fluorescence quenching was carried out by adding
increasing amounts of acrylamide (0–104 mm) to solutions
with the excitation wavelength set at 290 nm The effective
quenching constants were obtained from modified Stern–
(23 data points) [43]
GdmCl-induced unfolding and refolding
For equilibrium transition studies, protein samples (final
increasing concentrations of GdmCl (0–8 m) in 20 mm
20 h, the equilibrium was reached and intrinsic fluores-cence emission and far-UV CD spectra (0.2-cm cuvette)
reversibil-ity of the unfolding, protein samples were unfolded at
dithiothreitol and 100 lm EDTA, in the presence and
started by 20-fold dilution of the unfolding mixture, at
unfold-ing containunfold-ing decreasunfold-ing GdmCl concentrations The
which had been established as sufficient to reach equilib-rium, intrinsic fluorescence emission and far-UV CD
Data analysis
The changes in intrinsic fluorescence emission spectra at increasing GdmCl concentrations were quantified as the
according to
corre-sponding fluorescence intensity at that wavelength, respec-tively This quantity is an integral measurement, negligibly influenced by the noise, which reflects changes in the shape and position of the emission spectrum Far-UV CD and
analyzed by the singular value decomposition (SVD) algo-rithm [1,45] using the software matlab (MathWorks, South Natick, MA, USA)
SVD is useful to find the number of independent com-ponents in a set of spectra and to remove the high-fre-quency noise and the low-frehigh-fre-quency random error CD spectra in the 210–250 nm region or in the 250–310 region (0.2 nm sampling interval) were placed in a rectangular matrix A of n columns, one column for each spectrum collected in the titration The A matrix is decomposed by
where U and V are orthogonal matrices and S is a diago-nal matrix The columns of U matrix contain the basis spectra and the columns of the V matrix contain the
spec-trum Both U and V columns are arranged in terms of their decreasing order of the relative weight of