 |
INTRODUCTION |
Plants contain molecules that allow their
protection against pathogenic agents such as fungi and bacteria. Among
these molecules are purothionins, which are small basic globular
proteins found in the endosperm of wheat seeds (Triticum
aestivum) (Balls et al., 1942
). Three genetic variants exist,
namely 
1-, 2-, and 
-purothionin
(Fernandez de Caleya et al., 1976), all ~5 kDa and containing 45 amino acid residues. Due to the presence of several basic amino acid
residues, purothionins are positively charged at neutral pH. The
three-dimensional (3D) structure of purothionins has been solved by NMR
spectroscopy (Clore et al., 1986, 1987) and by x-ray diffraction (Rao
et al., 1995; Stec et al., 1995) and can be represented by the Greek
capital letter gamma (
). The vertical arm consists of two
antiparallel -helices and the horizontal arm contains a coil in
extended conformation and a short antiparallel -sheet. This tertiary
structure is stabilized by salt bridges, intramolecular hydrogen bonds,
and four disulfide bridges. Purothionins have an amphipathic character
with the hydrophobic residues located at the outer surface of the two
-helices, and the hydrophilic residues situated at the inner surface
of the and at the outer surface of the corner of the (Clore et
al., 1986; Teeter et al., 1981).
An important property of purothionins is their toxicity toward many
living organisms such as bacteria, yeast, and fungi (Stuart and Harris,
1942; Fernandez de Caleya et al., 1972), animals (Coulson et al.,
1942), cultured mammalian cells (Nakanishi et al., 1979; Carrasco et
al., 1981) and insect larvae (Kramer et al., 1979). Other biological
activities have been observed in vitro (for a review see Bohlmann and
Apel, 1991; Florack and Stiekema, 1994), but the main biological
function of purothionins appears to be related to the defense of plants
against their microbial predators. Most of the observed biological
activities result from the interaction of purothionins with the target
cell membrane (Carrasco et al., 1981). This is strengthened by the fact
that several thionins interact with phospholipid bilayers and natural
membranes (Caaveiro et al., 1997, 1998; Huang et al., 1997; Hughes et
al., 2000; Thevissen et al., 1996).
Huang et al. (1997) have studied the interaction between
Pyrularia thionin, an analog of purothionins, and large
unilamellar vesicles (LUV) of dipalmitoylphosphatidylglycerol (DPPG) by
differential scanning calorimetry (DSC). The results obtained indicated
that at low concentration of protein, the gel-to-fluid phase transition temperature of DPPG increases, in agreement with the formation of
electrostatic interactions between the thionin and the lipid vesicles.
However, at higher concentrations of protein, a decrease of the phase
transition temperature and the appearance of a second transition at a
lower temperature were observed. As the thionin concentration
increased, only the low-temperature phase transition was detected. This
low-temperature transition has been associated with the insertion in
the hydrophobic core of the DPPG bilayer of the tryptophan residue
located in the hydrophobic part of the -helices. This study
suggested that the interaction between the thionin and the lipid could
involve a two-step mechanism. First, the protein adsorbs to the bilayer
surface through electrostatic interactions, and second, penetrates the
hydrophobic bilayer region through a mechanism driven by the thionin tryptophan.
Caaveiro et al. (1997) have investigated the interaction between
-purothionin and large unilamellar vesicles composed of phosphatidylcholine (PC), phosphatidylglycerol (PG), and the mixture PC/PG (1:1). Their results indicate that -purothionin induces the
aggregation of the negatively charged vesicles and causes the release
of the fluorescent probes encapsulated in the vesicles. In addition,
these effects depend on the lipid composition and only occur when the
vesicles contain negatively charged phospholipids. From this study,
Caaveiro et al. have concluded that the protein induces membrane
leakage that causes the release of the intracellular material. They
have also studied the effect of the insertion of distearoylphosphatidylethanolamine-polyethylene glycol 2000 (PEG-PE) in
the bilayer on the aggregation and permeability of the vesicles (Caaveiro et al., 1998). Their results indicate that PEG-PE prevents the aggregation of the vesicles while it does not prevent the binding
of purothionin to the vesicles and, therefore, the release of
fluorescent probes. Hence, the presence of PEG-PE at the surface of the
membrane separates two processes that otherwise would be taking place
simultaneously, i.e., membrane permeabilization and aggregation.
More recently, it has been shown that ion channels are formed in model
lipid membranes composed of PE/PC 2:1 and PS/PE/PC 2:2:1 upon addition
of 1-, 2-, or -purothionin (Hughes et
al., 2000). However, other results have shown that fluorescent probes are released from vesicles in the presence of -purothionin even though these probes are too large to go through the ion channels formed
by the protein. This implies that the bilayer must necessarily be
destructured to allow the release of large molecules (Caaveiro et al.,
1998), in agreement with the fact that the -helices of purothionins
are too short to span a lipid bilayer (Caaveiro et al., 1998).
To better understand the mode of action of purothionins, we have
investigated the interaction between -purothionin and multilamellar vesicles (MLV) of dimyristoylphosphatidylglycerol, a negatively charged
phospholipid, using both 31P solid-state NMR and Fourier
transform infrared spectroscopy.
Solid-state 31P-NMR spectroscopy is a valuable technique to
study the different phases formed by model phospholipid membranes. The
spectra of the different lipid phases (e.g., gel and fluid lamellar
phases, inverted hexagonal phase, and isotropic phases such as small
vesicles and micelles) are characterized by a specific lineshape
(Seelig, 1978; Smith and Ekiel, 1984). Also, 31P-NMR allows
the study of the dynamics of the lipid headgroup. In the present study
we have investigated the lateral diffusion of the phospholipids in the
plane of the membrane using two-dimensional (2D) exchange spectroscopy
(Fenske and Jarrell, 1991). This technique uses the fact that in
solid-state NMR, the chemical shielding shows an orientational
dependence in the static magnetic field. Therefore, it is possible to
observe correlation peaks in a 2D spectrum map representing an
orientational exchange originating from the diffusion of the lipids
over the curved membrane surface. In this study, NMR spectroscopy has
been used to determine how -purothionin affects the organization of
DMPG membranes and the lateral diffusion of the phospholipids.
Fourier transform infrared spectroscopy (FTIR) is well-suited for the
study of lipid-protein interactions because it allows the investigation
of the conformation of phospholipid molecules at different levels in
the lipid bilayers and to follow structural changes that occur during
the gel-to-fluid phase transition. Moreover, it can provide information
on the secondary structure of proteins in lipid bilayers. Therefore, we
have used FTIR to study how -purothionin affects the acyl chains,
the interfacial region, and the phosphate groups of DMPG bilayers.
Furthermore, with the use of polarized infrared radiation, it has been
possible to determine the orientation of the protein relative to the bilayer.
| |
MATERIALS AND METHODS |
Materials
Dimyristoylphosphatidylglycerol was purchased from Avanti Polar
Lipids (Alabaster, AL) and used without further purification. Purothionins were extracted from wheat seeds using a modification of
the procedures previously described for the purification of puroindolines (Blochet et al., 1993; Dubreil et al., 1997). Briefly, 4 kg of wheat endosperm flour were extracted with 10 l of Tris 100 mM, pH 7.8, NaCl 0.1 M, EDTA 5 mM containing 5% Triton X-114. After
stirring for 12 h at 4°C and centrifugation at 8,000 × g for 30 min, the supernatant was heated at 30°C to allow
phase partitioning. The upper detergent-poor phase was discarded. The lower detergent-rich phase was recovered, diluted five times with water, and loaded on a column packed with a cation exchanger (SP Biobeads, Pharmacia, U.K.). As highlighted by SDS PAGE,
purothionins eluted as a single peak just after puroindolines and the
corresponding collected fractions were pooled and dyalized against
deionized water. After freeze-drying, purothionins were separated in
1, 2, and -purothionin by
semi-preparative reversed-phase HPLC on a column packed with Nucleosil
C18 5 µ 300 Å with a gradient from 0.1% TFA in deionized water to
0.1% TFA in CH3CN. The corresponding fractions were
collected and freeze-dried after dilution with deionized water. The
purity of the purified fractions was controlled by mass spectrometry as
previously described (Douliez et al., 2001).
Sample preparation
NMR experiments
Pure DMPG dispersions (20% w/v) were prepared by adding an
aqueous solution containing 20 mM NaCl and 1 mM EDTA to the solid lipids. The pH of the lipid dispersion was measured with a
microelectrode (Microelectrodes, Londonderry, NH) and adjusted to 7.0 with diluted NaOH or HCl. To obtain a homogenous dispersion, the sample
was heated at ~40°C for 5 min, stirred on a vortex mixer, and
cooled in liquid nitrogen. This cycle was repeated at least five times to obtain multilamellar vesicles. The lipid/protein complexes were
prepared by mixing the lipid dispersion (20 mg DMPG) and a protein
solution made in 20 mM NaCl and 1 mM EDTA. The protein concentration
was adjusted to yield the desired lipid-to-protein molar ratios and the
pH was verified and adjusted to 7.0 if necessary. The lipid-protein
mixtures were then heated (40°C), stirred, and cooled (10°) at
least five times.
FTIR experiments
The lipid dispersion (10%, w/v) and protein solution (2%, w/v)
were prepared in 150 mM NaCl made in H2O or
D2O. Deuterated water was used to eliminate the spectral
interference due to the bending vibration mode of water in the
1600-1700 cm
1 region. The pH of these samples was
measured and adjusted to 7.0 with diluted NaOH and HCl (or NaOD and
DCl). To obtain a homogenous DMPG dispersion, freeze-thawing was
applied as described above. The DMPG/-purothionin complexes were
prepared by mixing the appropriate volumes of lipid dispersion and
protein solution to obtain lipid-to-protein molar ratios of 30:1, 15:1,
and 10:1. The pH of the lipid-protein mixtures was verified and
adjusted to 7.0 if necessary. The complexes were formed by five
heating-cooling cycles as described above.
NMR measurements
The 31P-NMR spectra were acquired at 121.5 MHz on a
Bruker ASX-300 spectrometer (Bruker Canada Ltd., Milton, ON) operating
at a 1H frequency of 300.0 MHz. Experiments were carried
out with a broadband/1H dual-frequency 4-mm probehead.
One-dimensional spectra (2200 scans for DMPG and 7000 scans for the
complexes) were recorded using a Hahn echo pulse sequence under
conditions of proton decoupling during the acquisition (Rance and Byrd,
1983). The 90° pulse length was ~4.5 µs, the interpulse delay 
was 30 µs, and the recycle time was 4 s. A line-broadening of
200 Hz was applied to all spectra.
Two-dimensional spectra were acquired using the NOESY pulse sequence
with TPPI to give quadrature detection in both dimensions (Bodenhausen
et al., 1984):
|
(1)
|
where t1 is the evolution time,
tm is the mixing time, and
t2 is the detection period. The
tm were varied from 50 µs to 200 ms, the 90°
pulse length was ~4.5 µs, and the recycle time was 4 s. The
data sets contained 512 points in the F2 dimensions and 64 points in the F1 dimension, zero-filled to 512 to obtain a square matrix. Ninety-six scans were recorded for each serial file in
given 2D experiments. A line-broadening of 200 Hz was applied to all
spectra in the F2 dimension. For all 1D and 2D experiments,
the spectral width was 50 kHz and the chemical shifts (expressed in
ppm) were referenced relative to the signal of phosphoric acid at 0 ppm.
To determine the correlation time for lateral diffusion, the method
described by Picard et al. (1998) was used. This method is based on the
calculation of the time-dependent orientational autocorrelation
function of order 2, C2(tm):
|
(2)
|
where tm is the mixing time, 
is the
chemical shift, and 
1 and 2 are the
frequencies before and after the mixing time. Because both and are expressed in frequency units, the C2 values
are dimensionless. The autocorrelation function,
C2(tm), decays exponentially with
the mixing time and the decay factor is the correlation time for the
lateral diffusion, td:
|
(3)
|
FTIR measurements
The infrared spectra were recorded on a Nicolet Magna 550 (Thermo-Nicolet, Madison, WI) Fourier transform infrared spectrometer equipped with a narrow-band mercury-cadmium-telluride detector and a
germanium-coated KBr beam-splitter. For the recording of transmission
spectra, ~12 µl of sample was inserted between two BaF2
windows separated by a 6-µm Mylar spacer in a homemade cell thermoelectrically regulated (Pézolet et al., 1983). A total of
200 scans were averaged at a resolution of 2 cm1 and a
spectrum was recorded every 2°C between 0 and 50°C. For polarized
attenuated total internal reflectance (ATR) measurements, an ATR unit
(Harrick Scientific Co., Ossining, NY) and a parallelogram germanium
ATR crystal (50 × 20 × 2 mm, 45°) were used. Before use,
the crystal was first cleaned with chloroform and then in a plasma
cleaner sterilizer (Harrick Scientific Co.). Oriented films were
prepared by spreading ~30 µl of the aqueous sample on the germanium
ATR crystal. The sample was then dried and the crystal was placed in
the ATR unit. For each polarization (parallel (p) and perpendicular (s)
to the plane of incidence), a total of 250 scans were averaged at room temperature.
All spectral manipulations were performed with the Grams 386 software
(Galactic Industries Corporation, Salem, NH). The spectra in the
CH2 region (3030-2770 cm1) were
baseline-corrected using a cubic function. In the carbonyl and
phosphate regions (1680-1780 cm1 and 995-1320
cm1, respectively), the baseline correction was done with
a quadratic function. The spectra in the amide bands region (1380-1720
cm1) were baseline-corrected using a linear function in
the case of -purothionin and a quadratic function in the case of the
DMPG/-purothionin complexes. The position of the band due to the
CH2 symmetric C-H stretching mode was determined by
calculating the center of gravity of the bands at 85% of their height
(Cameron et al., 1982) and the carbonyl bands were deconvolved
according to the method of Kauppinen et al. (1981).
To determine the orientation of the -helices and -sheet of
-purothionin and of the DMPG acyl chains, the dichroic ratios RATR = (Ap/As) were calculated from
the height of the bands. For the DMPG acyl chains, this height was
measured at 2850 cm1 and for the -helices, at 1664 cm1. For the -sheet, the height was measured at 1635 and 1530 cm1 in the region of the amide I band and amide
II band, respectively.
For a system with both axial symmetry (e.g., acyl chains, -helices)
and uniaxial orientation, an order parameter
P2(cos 
)
with respect to the normal to
the ATR crystal can be determined with the following equation (Fringeli
and Günthard, 1981):
|
(4)
|
where RATR is the dichroic ratio, 
is
the angle between the transition moment of a given vibration and the
chain axis, and E
,
E
, and E
are the
mean-squared electric fields along the x, y, and
z directions, respectively. Values of 1.954, 2.303, and
2.651 along the x, y, and z directions,
respectively, were calculated for the mean-squared electric fields
using the Harrick thick film approximation (Harrick, 1967) with
refractive indices of 1.45 and 4.00 for the film and the germanium
crystal, respectively, and an angle of incidence of 45°. Angles of 90° and 38° have been used for the acyl chains in the
all-trans conformation and the -helices, respectively (Buffeteau et al., 2000; Marsh et al., 2000). Assuming an infinitely narrow distribution of orientation (Lafrance et al., 1995; Johansson and Lindblom, 1980), an angle can be determined from the
order parameter P2(cos ). A value of 1 for the order parameter P2(cos ) is
obtained for perfect orientation along the reference direction ( = 0°), while a value of 
0.5 is obtained for perfect
perpendicular orientation ( = 90°). Unordered structures or
structures perfectly oriented at 54.7° will yield a value of the
order parameter of 0.
Equation 4 cannot be used to calculate the orientation of a nonaxially
symmetric system such as a -sheet. To define the orientation of a
-sheet, two dichroic ratios are required. Marsh (1997) has developed
equations to determine the orientation of -sheets using the dichroic
ratio of both the amide I and amide II bands (Eqs. 5 and 6). This
method assumes that the resultant transition moment of a -sheet is
oriented parallel (
= 0°) to the -strand axis for the
amide II vibration and perpendicular ( = 90°) to this axis
for the amide I vibration. A combination of the dichroic ratios
obtained for both the amide I and amide II bands allows a complete
definition of the orientation of the -sheet.
|
(5)
|
|
(6)
|
In these equations, is the angle between the transition
moment and the chains axis, and E,
E, and E are the
mean-squared electric fields along the x, y, and
z directions, respectively. Equations 5 and 6 allow
calculation of the tilt angle, , of a -sheet relative to the
normal to the crystal and the tilt angle, , of the chains inside the
-sheet. It is considered that the angle is the same for all the
-strands constituting a -sheet.
| |
RESULTS AND DISCUSSION |
31P-NMR results
Spectral lineshapes
The phase behavior of aqueous dispersions of DMPG in the absence
and presence of -purothionin has first been investigated. The
31P-NMR spectra of pure DMPG and of DMPG/-purothionin
complexes at 10 and 40°C are shown in Fig.
1. As expected, the results show that the
spectral width decreases with increasing temperature as the lipid
undergoes the transition from the gel phase to the fluid phase. The
smaller spectral width obtained in the fluid phase can be explained by
an increased disordering of the lipid headgroup (Smith and Ekiel,
1984). In addition, the spectra obtained for both the pure DMPG and the
DMPG/-purothionin complexes at lipid-to-protein molar ratios of 30:1
and 15:1 are characteristic of a lamellar phase (Seelig, 1978). This
indicates that the presence of -purothionin does not induce the
formation of nonbilayer phases. Even though it has been suggested that
the protein disrupts the lipid bilayer (Carrasco et al., 1981; Huang et
al., 1997), the results obtained here do not show any evidence of the
formation of small isotropic vesicles (Picard et al., 1996). However,
the widths of the spectra obtained for the complex at a
lipid-to-protein molar ratio of 15:1 are slightly smaller than those
observed for the pure lipid, indicating that the presence of a large
proportion of protein could induce an increased disordering of the
lipid headgroup.
View larger version (18K):
[in this window]
[in a new window]
|
FIGURE 1
31P-NMR spectra obtained at 10°C
(left) and 40°C (right) for pure DMPG and for
DMPG/-purothionin complexes at lipid-to-protein molar ratios of 30:1
and 15:1.
|
|
The DMPG/-purothionin complex at a lipid-to-protein molar ratio of
10:1 has also been investigated, but it has been impossible in this
case to obtain an NMR spectrum with an acceptable signal-to-noise ratio. When the protein solution was added to the lipid dispersion, the
complex precipitated. This could be due to the aggregation of the
multilayered vesicles, resulting in a much broader NMR signal.
The results presented in Fig. 1 also indicate that the addition of the
protein causes a shift of the NMR spectra toward higher chemical shift
values in both the gel and fluid phases. This can be explained by the
nature of the interaction between -purothionin and DMPG. More
specifically, the presence of a positive charge near the negatively
charged phosphate groups decreases the electronic density around the
phosphorus nucleus, and consequently the spectrum is shifted toward
higher chemical shift values (Akitt, 1992). This behavior shows that
electrostatic interactions occur when -purothionin binds to DMPG.
Lateral diffusion
31P two-dimensional solid-state NMR spectroscopy
(Fenske and Jarrell, 1991) has been used to determine how
-purothionin affects the slow motions (<103 Hz) of the
lipids and more specifically, the lateral diffusion of DMPG. In
two-dimensional 31P-EXSY solid-state NMR spectra,
cross-intensity appears when there is a change of orientation during
the mixing time tm and results in a square
spectrum. In the case of spherical vesicles, two main motions
contribute to the change in the orientation of the phospholipids: the
tumbling, characterized by the correlation time
tt, and the lateral diffusion, characterized by
the correlation time td (Fenske and Jarrell,
1991). The correlation time measured experimentally, te, is related to the correlation time due to
the diffusion, td, and the correlation time due
to tumbling, tt (Eq. 7):
|
(7)
|
where
|
(8)
|
and
|
(9)
|
In these equations, 
is the viscosity of the solvent,
k is the Boltzmann constant, r is the radius of
the vesicles, T is the temperature, and
Dt is the diffusion constant.
In the fluid phase, the change in orientation is mainly due to the
lateral diffusion of the lipids. Indeed, since the mean calculated
radius for vesicles prepared by dispersion varies between 1.0 and 1.2 µm (Fenske and Jarrell, 1991), the correlation time due to tumbling
is ~0.8-1.4 s (Eq. 8). The td values for the
lateral diffusion being of the order of a millisecond,
1/td 
1/tt and thus,
te = td. Therefore,
tumbling can be neglected.
The lateral diffusion of pure DMPG and of DMPG/-purothionin
complexes at lipid-to-protein molar ratios of 30:1 and 15:1 has been
investigated. Fig. 2 shows the 2D spectra
obtained with mixing times of 50 µs and 5 ms for pure DMPG and the
two complexes. At short mixing time (Fig. 2, left), the
intensity is mainly located on the diagonal, indicating that there is
little or no lateral diffusion. At a longer mixing time (Fig. 2,
right), the off-diagonal spectral intensity becomes more
important, which indicates that the lipids change their orientation
during the mixing time and, therefore, that lateral diffusion occurs.
If we compare the spectra obtained for the complexes (Fig. 2,
middle and bottom) with those of pure DMPG (Fig.
2, top), little difference is seen. However, the calculation
of the autocorrelation time functions C2 as a function of mixing time (Fig. 3) clearly
indicates differences between the systems studied and allows
determination of the correlation time for the lateral diffusion,
td. The correlation time obtained for the pure
DMPG system is 4 ms and correlation times of 8 and 17 ms have been
obtained for the complexes at lipid-to-protein molar ratios of 30:1 and
15:1, respectively. These results indicate that -purothionin
decreases the lateral diffusion of the DMPG and therefore, that the
protein could act as an obstacle to the lipid diffusion.
View larger version (19K):
[in this window]
[in a new window]
|
FIGURE 2
2D 31P-NMR spectra of pure DMPG
(top) and of DMPG/-purothionin complexes at
lipid-to-protein molar ratios of 30:1 (middle) and 15:1
(bottom) at 40°C for mixing times of 50 µs and 5 ms.
|
|
View larger version (16K):
[in this window]
[in a new window]
|
FIGURE 3
Normalized C2 as a function of mixing time
for pure DMPG ( ) and for DMPG/-purothionin complexes at
lipid-to-protein molar ratios of 30:1 ( ) and 15:1 ( ).
|
|
Infrared spectroscopy results
Lipid membrane structure
To determine how -purothionin interacts with DMPG, we have
investigated the effects of the protein on the acyl chains, the interfacial region (C
O), and the polar headgroup
(PO2) of DMPG bilayers by infrared spectroscopy.
To evaluate the effect of -purothionin on the lipid acyl chains, we
have investigated the symmetric CH2 stretching mode (2850 cm1) because it is almost unaffected by underlying
contributions from the protein component (Surewicz et al., 1987). The
evolution of the frequency of the symmetric methylene stretching mode
of the pure DMPG acyl chains as a function of temperature shows a discontinuity at the gel-to-fluid transition temperature
(Tm) (Fig. 4), due
to the important increase in the proportion of gauche conformers at Tm (Mendelsohn and Mantsch, 1986;
Lewis and McElhaney, 1998). From these results, a phase transition
temperature of 25°C is obtained for pure DMPG, in agreement with the
value of 24°C found in the literature (Marsh, 1990). The addition of
-purothionin at a lipid-to-protein molar ratio of 30:1 shifts the
phase transition temperature to 32°C. For the lipid-to-protein molar
ratio of 15:1, the main phase transition is observed near 31°C, but a
second transition is observed at 18°C. For the complex at a
lipid-to-protein molar ratio of 10:1, the high-temperature phase
transition occurs near 30°C and the low-temperature transition is
further shifted to 16°C (Fig. 4).
View larger version (18K):
[in this window]
[in a new window]
|
FIGURE 4
Temperature dependence of the CH2 symmetric
stretching vibration of pure DMPG () and of DMPG/-purothionin
complexes at lipid-to-protein molar ratios of 30:1 (), 15:1 ( ),
and 10:1 ( ).
|
|
An increase of the phase transition temperature of phospholipids is
generally observed in the case of electrostatic interactions with
proteins and polypeptides (Paphadjopoulos et al., 1975; Carrier et al.,
1997). More specifically, it has been shown that an increase of
Tm is associated with the stabilization of the
charges between the protein and the lipid, leading to an increase of
the order of the acyl chains and to the stabilization of the gel phase
in the presence of protein (Surewicz and Epand, 1986). The results presented in Fig. 4 also indicate that the acyl chains are perturbed by
the presence of the protein. Indeed, for the complex at a
lipid-to-protein molar ratio of 10:1 in the gel phase and for all the
complexes in the fluid phase, the wavenumbers of the methylene
stretching band are higher in the lipid-protein systems than those
observed for the pure DMPG bilayers. This observation shows that the
presence of -purothionin results in a decrease of the conformational
order of the lipid acyl chains, suggesting the insertion of the protein in the lipid bilayer in addition to the electrostatic interaction discussed above.
The presence of a second phase transition at lower temperature for the
complexes with lipid-to-protein molar ratios of 15:1 and 10:1 also
supports the existence of hydrophobic interactions between the protein
and the membrane. Indeed, by inserting into the bilayer, the
hydrophobic residues of the protein would cause a decrease of the acyl
chain order. Thus, the phospholipids in direct contact with the protein
would have a phase transition temperature lower than that of the lipids
that do not interact directly with the protein. This thermotropic
behavior is similar to that observed by Huang et al. (1997) in a DSC
study on the interaction of the thionin of Pyrularia pubera
and dipalmitoylphosphatidylglycerol (DPPG) vesicles. In this study, the
decrease of the high-temperature phase transition and the appearance of
a low-temperature phase transition have been associated with the
insertion in the hydrophobic core of the lipid bilayer of the
tryptophan residue, located in the hydrophobic part of the -helices.
Our results suggest that the binding of -purothionin to DMPG
membranes is similar to that of the Pyrularia pubera thionin
to DPPG. However, a tyrosine residue also located in the hydrophobic
part of the -helices would insert in the bilayer instead of a
tryptophan. The strong electrostatic and hydrophobic interactions
between -purothionin and DMPG suggested by the infrared results are
in agreement with the decrease of the lipid lateral diffusion observed
by 2D 31P-NMR spectroscopy.
The interfacial region of the phospholipid bilayers has also been
investigated using the carbonyl stretching vibration between 1700 and
1750 cm1 (Casal and Mantsch, 1984). Only one band is
generally observed in this spectral region for hydrated samples, but
the Fourier deconvolution technique (Kauppinen et al., 1981) allows
enhancing the resolution and revealing two overlapping bands. The band
at 1725 cm1 is associated with carbonyl groups involved
in hydrogen bonds, while the band at 1740 cm1 is
associated with free carbonyl groups (Blume et al., 1988). The
intensity ratio of these two bands is therefore useful for monitoring
the hydration of the interfacial region of a lipid bilayer.
Fig. 5 shows the deconvolved spectra in
the carbonyl region at 10 and 40°C for DMPG in D2O in the
absence and in the presence of -purothionin. For pure DMPG, the
relative intensity of the hydrogen-bonded carbonyl component (1725 cm1) increases when the lipids undergo the gel-to-fluid
phase transition. This behavior can be explained by the fact that the
hydrocarbon chains become more disordered at high temperature,
facilitating the penetration of water in the interfacial region of the
bilayer. The proportion of CO groups bonded to water molecules is
therefore more important in the fluid phase. The results presented in
Fig. 5 also indicate that the bonded CO band is shifted from 1723 cm1 to 1728 cm1 from the gel to the fluid
phase. This can be explained by the fact that at high temperature, the
hydrogen bond between the lipid CO group and the deuterium atoms of
the solvent are weaker than at low temperature and, consequently, this
band is shifted toward higher wavenumbers (Carrier et al., 1997).
View larger version (19K):
[in this window]
[in a new window]
|
FIGURE 5
Infrared spectra in the lipid carbonyl stretching mode
region for pure DMPG (solid line) and for
DMPG/-purothionin complexes at lipid-to-protein molar ratios of 30:1
(dotted line) and 15:1 (dot-dashed line).
|
|
For the lipid/protein complexes, the results indicate that the relative
intensity of the hydrogen-bonded carbonyl component increases compared
to that of pure DMPG and this, in both the gel and fluid phases. Two
hypotheses can be suggested. On one hand, -purothionin could favor
the penetration of water in the interfacial region or, on the other
hand, hydrogen bonds could be formed between the carbonyl groups of
DMPG and the amino acid side chains of -purothionin. It has been
shown that the increase of the relative intensity of the
hydrogen-bonded carbonyl component is generally associated with the
formation of hydrogen bonds between CO groups and protons (or
deuterium atoms) of the solvent (Désormeaux et al., 1992; Carrier
et al., 1997; Arrondo and Goni, 1998; Nabet et al., 1994). Therefore,
our results are consistent with the hypothesis that -purothionin
would modify the DMPG bilayer sufficiently to increase the
accessibility of the water molecules in the interfacial region, as also
suggested by Huang et al. (1997). However, the existence of hydrogen
bonds between the lipid carbonyl groups and the -purothionin amino
acid side chains cannot be ruled out from our results.
The study of the phosphate group absorption bands between 900 and 1300 cm1 (Fringeli and Günthard, 1981) allows
characterizing the polar headgroups of phospholipids (Arrondo and Goni,
1998). More specifically, a perturbation of the phosphate bands
suggests the presence of an interaction between a protein and the
membrane surface. The results obtained in the present study (not shown)
indicate that the phosphate bands of DMPG are not significantly
perturbed by the presence of -purothionin. This behavior is in
agreement with that observed in the study of the interaction between
DMPG and other basic proteins (Surewicz et al., 1987).
Orientation measurements
The orientation of the -helices and the -sheet of
-purothionin for lipid/protein complexes at lipid-to-protein molar
ratios of 30:1, 15:1, and 10:1 and the orientation of the lipid acyl chains in pure DMPG and in all the complexes has also been
investigated. Fig. 6 shows the polarized
ATR spectra of the DMPG/-purothionin complex (30:1) in the CH and
amide I and II regions. In addition, Table
1 presents the dichroic ratios, the order
parameters, and the tilt angles of the -helices, -sheet, and
lipid acyl chains. These values are the average of three independent
measurements. The tilt angles have been calculated assuming an
infinitely narrow distribution of orientation (Lafrance et al., 1995;
Johansson and Lindblom, 1980).
View larger version (20K):
[in this window]
[in a new window]
|
FIGURE 6
S (dotted line) and p-polarized (solid
line) ATR spectra in the C-H stretching region (top)
and in the amide I and amide II spectral region (bottom)
obtained in the fluid phase (40°C) for the DMPG/-purothionin
complex at a lipid-to-protein molar ratio of 30:1.
|
|
View this table:
[in this window]
[in a new window]
|
TABLE 1
Orientation of the -helices, the -sheet, and the
lipid acyl chains in DMPG/-purothinin complexes as a function of the
lipid-to-protein molar ratio
|
|
The results presented in Table 1 indicate that the -helices have an
order parameter P2(cos ) between 0.08
and 0.16 as a function of the lipid-to-protein molar ratios,
corresponding to tilt angles between 58° and 61° relative to the
crystal normal. Within experimental errors, these tilt angles do not
vary significantly as a function of the lipid-to-protein molar ratio.
However, an order parameter P2(cos ) of
0.63 was obtained for the -sheet, corresponding to a tilt angle of
30°. Although the angle obtained for the -helices is close to
54.7°, the angle of an isotropic distribution, the combination of
tilt angles obtained for both the -helices and the -sheet, is
consistent with an orientation of -purothionin relative to the DMPG
membrane as illustrated in Fig. 7.
The acyl chains of pure DMPG have an order parameter
P2(cos ) of 0.72, corresponding to an
orientation of 26° relative to the crystal normal, assuming an
infinitely narrow distribution of orientation. This is in close
agreement with the angle of 29° obtained by x-ray diffraction
(Pascher et al., 1987). When -purothionin is added at
lipid-to-protein molar ratios of 30:1, 15:1, and 10:1, the order
parameter P2(cos ) considerably
decreases from 0.72 for pure DMPG to values between 0.48 and 0.58 in
the complexes, indicating an increased conformational disorder of the
lipid acyl chains. This behavior is in agreement with the partial
insertion of the protein in the bilayer, as suggested above. Within
experimental error, no significant differences are observed between the
different lipid-to-protein molar ratios.
| |
CONCLUSIONS |
The interaction between DMPG bilayers and -purothionin has been
investigated in the present study by 31P-NMR and infrared
spectroscopy. The 31P-NMR results indicate that the
organization of the lipid bilayer is not significantly affected by the
presence of -purothionin and that the protein decreases the lateral
diffusion of DMPG. However, the results obtained by infrared
spectroscopy demonstrate the existence of an electrostatic interaction
between DMPG and -purothionin and a partial insertion of the protein
in the lipid membrane. Orientation measurements are consistent with a
model in which the -helices and the -sheet have tilt angles of
~60° and 30°, respectively, relative to the DMPG membrane.
This work was supported by the Natural Science and Engineering
Research Council (NSERC) of Canada, by the Fonds pour la Formation de
Chercheurs et pour l'Aide à la Recherche (FCAR) from the
province of Québec, and by the Centre de Recherche en Sciences et
Ingénierie des Macromolécules (CERSIM). J.-A.R. also thanks
NSERC for the award of a postgraduate scholarship.
Address reprint requests to Michèle Auger, Département de
Chimie, CERSIM, Université Laval, Québec G1K7P4, Canada.
Tel.: 418-656-3393; Fax: 418-656-7916; E-mail:
michele.auger{at}chm.ulaval.ca.
-Purothionin and
Dimyristoylphosphatidylglycerol: A 31P-NMR and Infrared
Spectroscopic Study
TOP
ABSTRACT
INTRODUCTION
