Article Information

• PDF (201 kb)
• Full Text with Large Figures

PubMed

• Articles by Frédéric Picard
• Articles by Michèle Auger

Related Articles

• Magnetically Aligned Phospholipid Bilayers with Positive Ord...

  Magnetically Aligned Phospholipid Bilayers with Positive Ordering: A New Model Membrane System Biophysical Journal, Volume 74, Issue 5, 1 May 1998, Pages 2405-2418 R. Scott Prosser, Jimmy S. Hwang and Regitze R. Vold Abstract A stable smectic phospholipid bilayer phase aligned with the director parallel to the magnetic field can be generated by the addition of certain trivalent paramagnetic lanthanide ions to a bicellar solution of dimyristoylphosphatidylcholine (DMPC) and dihexanoylphosphatidylcholine (DHPC) in water. Suitable lanthanide ions are those with positive anisotropy of their magnetic susceptibility, namely Eu, Er, Tm, and Yb. For samples doped with Tm, this phase extends over a wide range of Tm concentrations (6–40mM) and temperatures (35–90°C) and appears to undergo a transition from a fluid nematic discotic to a fluid, but highly ordered, smectic phase at a temperature that depends on the thulium concentration. As a membrane mimetic, these new, positively ordered phospholipid phases have high potential for structural studies using a variety of techniques such as magnetic resonance (EMR and NMR), small-angle x-ray and neutron diffraction, as well as optical and infrared spectroscopy. Abstract | Full Text | PDF (281 kb)

• Insights on the Interactions of Synthetic Amphipathic Peptid...

  Insights on the Interactions of Synthetic Amphipathic Peptides with Model Membranes as Revealed by P and H Solid-State NMR and Infrared Spectroscopies Biophysical Journal, Volume 90, Issue 11, 1 June 2006, Pages 4071-4084 Marise Ouellet, Geneviève Bernard, Normand Voyer and Michèle Auger Abstract We studied the interaction between synthetic amphipathic peptides and model membranes by solid-state NMR and infrared spectroscopies. Peptides with 14 and 21 amino acids composed of leucines and phenylalanines modified by the addition of crown ethers were synthesized. The 14-mer and 21-mer peptides both possess a helical amphipathic structure. To shed light on their membrane interaction, P and H solid-state NMR experiments were performed on both peptides in interaction with dimyristoylphosphatidylcholine vesicles in the absence and presence of cholesterol, dimyristoylphosphatidylglycerol vesicles, and oriented bicelles. P NMR experiments on multilamellar vesicles reveal that the dynamics and/or orientation of the polar headgroups are weakly yet markedly affected by the presence of the peptides, whereas P NMR experiments on bicelles indicate no significant changes in the morphology and orientation of the bicelles. On the other hand, H NMR experiments on vesicles reveal that the acyl chain order is affected differently depending on the membrane lipidic composition and on the peptide hydrophobic length. Finally, infrared spectroscopy was used to study the interfacial region of the bilayer. Based on these studies, mechanisms of membrane perturbation are proposed for the 14-mer and 21-mer peptides in interaction with model membranes depending on the bilayer composition and peptide length. Abstract | Full Text | PDF (233 kb)

• Solid-State NMR Studies of a Diverged Microsomal Amino-Proxi...

  Solid-State NMR Studies of a Diverged Microsomal Amino-Proximate Δ12 Desaturase Peptide Reveal Causes of Stability in Bilayer: Tyrosine Anchoring and Arginine Snorkeling Biophysical Journal, Volume 90, Issue 4, 15 February 2006, Pages 1249-1259 William J. Gibbons, Ethan S. Karp, Nick A. Cellar, Robert E. Minto and Gary A. Lorigan Abstract This study reports the solid-state NMR spectroscopic characterization of the amino-proximate transmembrane domain (TM-A) of a diverged microsomal Δ12-desaturase (CREP-1) in a phospholipid bilayer. A series of TM-A peptides were synthesized with H-labeled side chains (Ala-53, -56, and -63, Leu-62, Val-50), and their dynamic properties were studied in 1,2-dimyristoyl--glycero-3-phosphatidylcholine (DMPC) bilayers at various temperatures. At 6mol % peptide to lipid, P NMR spectra indicated that the peptides did not significantly disrupt the phospholipid bilayer in the L phase. The H NMR spectra from Ala-53 and Ala-56 samples revealed broad Pake patterns with quadrupolar splittings of 16.9kHz and 13.3kHz, respectively, indicating restricted motion confined within the hydrocarbon core of the phospholipid bilayer. Conversely, the deuterated Ala-63 sample revealed a peak centered at 0kHz with a linewidth of 1.9kHz, indicating increased side-chain motion and solvent exposure relative to the spectra of the other Ala residues. Val-50 and Leu-62 showed Pake patterns, with quadrupolar splittings of 3.5kHz and 3.7kHz, respectively, intermediate to Ala-53/Ala-56 and Ala-63. This indicates partial motional averaging and supports a model with the Val and Leu residues embedded inside the lipid bilayer. Solid-state NMR spectroscopy performed on the H-labeled Ala-56 TM-A peptide incorporated into magnetically aligned phospholipid bilayers indicated that the peptide is tilted 8° with respect to the membrane normal of the lipid bilayer. Snorkeling and anchoring interactions of Arg-44 and Tyr-60, respectively, with the polar region or polar hydrophobic interface of the lipid bilayer are suggested as control elements for insertional depth and orientation of the helix in the lipid matrix. Thus, this study defines the location of key residues in TM-A with respect to the lipid bilayer, describes the conformation of TM-A in a biomembrane mimic, presents a peptide-bilayer model useful in the consideration of local protein folding in the microsomal desaturases, and presents a model of arginine and tyrosine control of transmembrane protein stability and insertion. Abstract | Full Text | PDF (334 kb)

• …more

Copyright © 1999 The Biophysical Society. All rights reserved.
Biophysical Journal, Volume 77, Issue 2, 888-902, 1 August 1999

doi:10.1016/S0006-3495(99)76940-7

Membranes

³¹P NMR First Spectral Moment Study of the Partial Magnetic Orientation of Phospholipid Membranes

Frédéric Picard, Marie-Josée Paquet, Jérôme Levesque, Anne Bélanger and Michèle Auger^,  

Département de Chimie, Centre de Recherche en Sciences et Ingénierie des Macromolécules, Université Laval, Québec, Québec G1K 7P4, Canada

Address reprint requests to Michèle Auger, Département de Chimie, CERSIM, Université Laval, Québec, Québec G1K 7P4, Canada. Tel.: 418-656-3393; Fax: 418-656-7916.

The spontaneous orientation of phospholipid membranes in magnetic fields has often been considered problematic due to the resulting change in the solid-state nuclear magnetic resonance (NMR) spectral lineshapes (Bayerl and Bloom, 1990; Van Etcheld et al., 1982; Killian and de Kruijff, 1986). However, this phenomenon is now widely exploited since it has been shown to improve the spectral resolution (Sanders and Prestegard, 1990,Vold and Prosser, 1996). Phospholipids spontaneously form unilamellar and multilamellar vesicles, which are usually spherical in an aqueous medium (Helfrich, 1973,Pidgeon et al), but some membranes, depending on composition, concentration, temperature, and the method of preparation, tend to modify their shape in the presence of a magnetic field (Seelig et al,Speyer et al,Jansson et al,Qiu et al). When sterols (Brumm et al,Reinl et al) or proteins (Neugebauer et al,Van Echteld et al,Pott and Dufourc, 1995) are embedded in phospholipid membranes, the orientation may be enhanced or modified.

The origin of the magnetic alignment for a given molecule is its diamagnetic or paramagnetic anisotropy (Boroske and Helfrich, 1978). If a diamagnetic molecule is placed in a magnetic field, a magnetic moment proportional to the diamagnetic anisotropy is induced. Most of the time, the magnetic energy of one molecule is not sufficient to allow its alignment. However, when several molecules are packed together, there may be a reorientation of the aggregate relative to the field (Qiu et al). The diamagnetic anisotropy of an axially symmetric molecule, Δχ, is defined by the difference between the diamagnetic susceptibility parallel (χ_‖) and perpendicular (χ_⊥) to the main axis of the molecule. For molecules with a negative Δχ, a perpendicular alignment will occur whereas those with a positive Δχ will orient parallel to the field. For example, in phospholipids, the hydrocarbon chains, which are considered to be the reference axis, have a Δχ<0 (Boroske and Helfrich, 1978), whereas glycerol ester carbonyls have a Δχ>0 (Maret and Dransfeld, 1985). Because glycerol carbonyls are about perpendicular to the phospholipid hydrocarbon chains (Sakurai et al), they both tend to orient the long axis of the lipid at 90° relative to the magnetic field. Peptide bonds also have a positive Δχ (Worcester, 1978). Because all peptide bonds contained in α or β helices are parallel to the helix axis, helical proteins will tend to orient parallel to a magnetic field (Neugebauer et al).

Proteins can also contain aromatic residues that are known to have large negative Δχ (Maret and Dransfeld, 1985). In particular, gramicidin A (GA) has 4 tryptophan residues in which the planes of the aromatic rings are approximately parallel to the helix axis (Ketchem et al), favoring a parallel orientation. In the same way, an aromatic molecule such as adriamycin (ADM) is expected, due to its negative Δχ, to orient its plane parallel relative to a magnetic field. In fact, it has been shown that the anthraquinone moieties of ADM molecules are closely stacked together when they are present at high concentration on the surface of a negatively charged phospholipid membrane (Goormaghtigh and Ruysschaert, 1984,De Wolf et al). Their magnetic susceptibilities are therefore added up, which makes possible the orientation of the membrane. The orientation of this membrane will depend on the angle formed by the ADM and the normal to the bilayer. If the ADM molecule is tilted by 39° (Goormaghtigh et al), the membrane would orient its normal parallel to the magnetic field.

The shape of phospholipidic membranes under high magnetic field is usually prolate ellipsoid (Helfrich, 1973,Speyer et al). These structures have been observed by electron microscopy (Brumm et al), ³¹P NMR (Pott and Dufourc, 1995,Brumm et al) and ²H NMR spectroscopies (Reinl et al,Schaefer et al). In addition, specific phospholipid mixtures are known to form bicelles, which are discoidal micelles formed when a long chain lipid (such as dimyristoylphosphatidylcholine (DMPC)) is mixed either with a short chain lipid (such as dihexanoylphosphatidylcholine (DHPC)) (Sanders and Schwonek, 1992,Vold and Prosser, 1996,Sanders and Prosser, 1998), a bile salt such as sodium glycocholate (Ram and Prestegard, 1988), or the detergent 3-(cholamidopropyl)dimethylammonio-2-hydroxy-1-propanesulfonate (CHAPSO) (Sanders and Prestegard, 1990,Hare et al). In addition, it was recently observed that the addition of small amounts of paramagnetic ions to pure bicelles results in systems in which the director is oriented parallel to the magnetic field (Prosser et al). This phenomenon is due to paramagnetism and will not be considered here.

Solid-state ³¹P NMR spectroscopy is an ideal technique to follow the orientation of phospholipid membranes due to the presence of only one phosphorus atom in each phospholipid and the absence of phosphorus in most proteins. In addition, several interactions studied by ³¹P NMR are orientation dependent, such as the dipolar coupling and the chemical shift anisotropy (CSA) (Seelig, 1978,Smith and Ekiel, 1984). Spectral simulations have been done to extract the ratio of ellipsoid long/short axis from experimental ³¹P (Pott and Dufourc, 1995,Brumm et al) and ²H NMR spectra (Reinl et al,Schaefer et al). In contrast, even though ³¹P and ²H NMR spectroscopies have been applied to bicelles, no spectral simulation of such system has been done to our knowledge. However, the shape of the bicelles has been investigated by Chung and Prestegard, 1993 based on field gradient studies and by Vold and Prosser, 1996 based on the ratios of the deuteron splittings for DHPC and DMPC.

In the present paper, a method is proposed that uses the first spectral moment of ³¹P NMR spectra to obtain structural details on magnetically oriented phospholipid membranes. More specifically, it can be used to determine either the dimension of a bicelle or the ratio of an ellipsoid long/short axis. All these parameters are extracted from equations that are derived in the Theory section. Different systems are studied to present several of the possibilities mentioned above. First, DMPC forms ellipsoidal vesicles in which the phospholipids have a perpendicular orientation relative to the magnetic field. Moreover, a DMPC:DHPC mixture gives bicelles that align the bilayer normal perpendicular to the magnetic field. Finally, the lipids in membranes made of DMPC:DHPC and GA have a parallel orientation relative to the field, such as those in a complex made of a mixture of cardiolipin (CL) and ADM.

In axially symmetric systems, ³¹P NMR chemical shifts are defined as

(1)

For static samples, a relationship between the spectral density, S(ω), and the angular distribution, P(β), called the principle of differential conservation of the integral (Schmidt-Rohr and Spiess, 1994) is given by

(2)

(3)

The angular distribution for ellipsoids has already been investigated by Pott and Dufourc, 1995. Our approach is very similar to theirs but the following presentation is still necessary for a better understanding of that used for bicelles. Also, the description of the origin of the ellipsoidal angular distribution will help to define the effect of partial orientation on the first spectral moment.

Ellipsoids can be represented by the parametric equations,

(4)

Display large version of this figure

Figure 1

Geometrical representations of (A) an ellipsoid, (B) a bicelle in the x′–y′ plane, and (C) a bicelle in the (x′y′)–z′ plane.

The tangents to the surface of the ellipsoid, _θ and _φ, can be obtained by the partial derivation of Eq. (4) relative to φ and θ,

(5)

(6)

(7)

(8)

(9)

(10)

(11)

(12)

In this model, the outer part of a bicelle is a torus and the inner part is a circular plane. A representation of a bicelle can be seen in Fig. 1. These two parts of the bicelle will be treated separately. The parametric equations of a torus are

(13)

(14)

(15)

(16)

(17)

(18)

(19)

(20)

(21)

(22)

(23)

(24)

(25)

(26)

As discussed above, if there is a change in the shape of the orientation distribution, there is also a change in the weighted-average frequency. Rearranging Eq. (23), it is possible to define a distribution order parameter,

(27)

(28)

(29A)

(29B)

(29C)

(30)

DMPC, DHPC, and CL were obtained from Avanti Polar Lipids (Alabaster, AL) and used without further purification. GA and ADM (doxorubicin hydrochloride) were obtained from Fluka (Ronkonkoma, NY) and used without further purification.

Aqueous dispersions of DMPC and CL were prepared in a 150mM NaCl and 10mM EDTA solution and adjusted to pH 6.5. Samples containing 20% (wt/wt) of lipids were then heated to ∼50°C for 10min, stirred on a vortex mixer, and cooled down at 0°C for 10min. This cycle was repeated at least five times just before the analysis. The solution of ADM (0.01g/mL) was prepared in 150mM NaCl and 10mM EDTA and adjusted to pH 6.5. The appropriate volume of the ADM solution was added to the CL dispersion to obtain a 2:1 CL:ADM molar ratio and then, five freeze–thaw cycles were applied to the system. Samples of 2.7:1 DMPC:DHPC molar ratio were prepared in a 0.1M KCl buffer. Then samples containing 20% (wt/wt) of lipids were heated to ∼35°C for 10min, stirred on a vortex mixer, sonicated, and cooled down at 0°C for 10min. This cycle was repeated at least five times just before the analysis. Samples of 10:4.3:1 DMPC:DHPC:GA molar ratio were prepared in trifluoroethanol. To obtain homogeneous peptide/lipid systems, the samples were incubated at 52°C for 1h and shaken on a vortex mixer at least a few times during the incubation cycle. After the incubation, the organic solvent was evaporated with a nitrogen stream followed by high vacuum pumping overnight. The samples were then hydrated at 32% (wt/wt) with a HEPES buffer prepared at pH 7.0 and were submitted to several cycles of heating (52°C), vortex-mixer, sonication, and cooling (0°C).

The ³¹P NMR spectra were acquired at 121.5MHz on a Bruker ASX-300 (Bruker Canada Ltd., Milton, ON) operating at a ¹H frequency of 300.0MHz. Experiments were carried out with a broadband/¹H dual frequency 4-mm probehead. The free induction decays (2K data points) were recorded with a spin echo sequence (2000 or 4000 scans) with a 4 to 7s repetition time and under conditions of proton decoupling. The ¹H 90° pulse length was typically 4.0μs, corresponding to a rotating-frame frequency of about 63kHz. The ³¹P 90° pulse length was 5μs and the interpulse delay was set to 30μs to avoid anisotropic T₂ effects on static spectra. The temperature was controlled to within ±0.5°C and the chemical shifts expressed in parts per millions (ppm) were referenced relative to the signal of phosphoric acid at 0ppm. When not specified, a line broadening of 50Hz was applied to the spectra. Magic angle spinning (MAS) ³¹P NMR spectra were obtained with a spinning speed of 6kHz and a free induction decay of 4K data points (1000 scans). No line broadening was applied to these spectra.

Simulations and calculations were performed with the Grams 386 software (Galactic Industries Corp., Salem, NH) using the Array Basic programming language. The simulated spectra were broadened by convolution with a Gaussian function. The experimental spectral treatment was done using the UXNMR software from Bruker (Bruker Canada Ltd., Milton, ON).

In this section, we will present the simulated ³¹P NMR spectra for partially oriented membranes, using the formalism developed in the Theory section. The simulated spectra will then be used to obtain the order parameters S₁ and S₂, respectively related to the orientation and the dynamics of membranes.

Figure 1 shows two types of partially oriented membranes, an ellipsoidal vesicle in A and two views of a bicelle in B and C. If partial orientation occurs in model membranes, they will most likely adopt an ellipsoidal structure in which a will be smaller than c due to the sign of the lipid diamagnetic anisotropy (Boroske and Helfrich, 1978). In this kind of system, the majority of the phospholipids will be oriented with their main axis, , at 90° relative to the surface. Such an ellipsoidal orientation implies either very large unilamellar or multilamellar vesicles because phospholipids in an ellipsoid with a very short semi-axis a will experience a fast angular diffusion that would result in chemical shifts characteristic of a hexagonal phase rather than the ones characteristic of lamellar phases (Seelig, 1978).

Figure 2 shows the simulated spectra of ellipsoidal structures for r ranging from 5 to 0.2. At a ratio of 5, the phospholipids are almost all oriented at 90° relative to the magnetic field. The ratio of 1 corresponds to a spherical vesicle, and, at a ratio of 0.2, the majority of the phospholipids is oriented parallel to the magnetic field. The surprising point of these simulated spectra is that a prolate ellipsoid gives a very small distribution of frequencies in comparison with the broad distribution occurring for an oblate shape. This can be seen in Fig. 2 by comparing the spectra with r=2.5 and 0.4.

Display large version of this figure

Figure 2

³¹P NMR spectral simulations of ellipsoidal vesicles as a function of r with δ=25ppm, δ_iso=0ppm and a Gaussian broadening (FWHH=1ppm).

The second case of partial orientation is the bicelle, which is usually made of a long acyl chain phospholipid that has been assumed to be on the two circular planes, and of a short acyl chain phospholipid that has been assumed to be on the rim of the bicelle (Sanders and Schwonek, 1992,Vold and Prosser, 1996). Bicelles made of only two circular planes have also been proposed for systems composed of CHAPSO and DMPC (Sanders and Prestegard, 1990,Sanders et al). In the present study, the bicelles are modeled by the combination of the outer part of a torus and of two circular planes (Figure 1BC). Simulations will be made using two different models. In the first model, we will consider a constant molecular surface area, a constant composition of phospholipids over the whole surface of the bicelle and a negligible contribution of the lipid lateral diffusion. In the second model, we will consider a rapid axial rotation of the bicelles around their main axis in addition to a rapid lateral diffusion of the lipids. In these two models, the bicelles are considered to be made of a single bilayer in which the thickness is the double of the radius of the edge, as illustrated in Figure 1C. The axis system of the bicelles is denoted (x′, y′, z′) to account for the difference with the laboratory frame (x, y, z). Two different orientations of the main axis of the bicelles are modeled, a bicelle with its z′ axis along z and a bicelle with its x′ axis along z.

Figure 3A reports the simulation of ³¹P NMR spectra of static parallel bicelles in which the effect of the orientation can be easily seen as an increasing spectral component at δ_‖. An important point to note is that a highly oriented system, e.g., with an r value of 0.5 corresponding to a bicelle thickness of 50Å and to a bicelle total diameter of 200Å, gives a broad spectrum if the main axis of the bicelle is oriented parallel to the magnetic field. However, because the broad component is due to the rim of the bicelle, static parallel bicelles should give sharp NMR peaks for molecules exhibiting fast axially symmetric motions inserted into their flat sections, such as membrane proteins. Finally, the spectral difference between parallel bicelles and oblate spheroids allows a discrimination of the two types of organization, as discussed in a later section.

Display large version of this figure

Figure 3

³¹P NMR spectral simulations of bicelles as a function of r with δ=25ppm, δ_iso=0ppm and a Gaussian broadening (FWHH=3ppm). (A) Static parallel bicelles. (B) Parallel bicelles with axial tumbling and lateral diffusion. (C) Perpendicular bicelles with axial tumbling and lateral diffusion.

Figure 3BC, shows simulated parallel and perpendicular bicelle spectra in which tumbling and lateral diffusion have been considered. In this model, we supposed that the phospholipids that constitute the torus do not diffuse in the planes and vice-versa. This assumption has been made because experimental bicelles are constituted of two types of phospholipids that are believed to be laterally phase separated. The relative proportion of the two peaks observed in these spectra varies with the r shape parameter. This relationship is defined by Eq. (22) and is plotted in Figure 4A. Therefore, from an experimental spectrum, it is possible to determine the size of the bicelle only by integrating the two peaks. The frequency of the peak from the torus part can also be related to the r shape parameter by

(31)

Display large version of this figure

Figure 4

(A) Relative proportion, ρ, and S_dist for the lipids in the torus part of the bicelles. (B) S₁ order parameter for ellipsoidal vesicles, parallel bicelles, and perpendicular bicelles as a function of the shape parameter r.

As discussed before, in oriented systems, it is possible to define an order parameter from the first spectral moment measured experimentally, which will be a measure of the orientation level. A system with a value of S₁ close to 1 (−0.5) indicates that all the molecules are oriented with their main axis parallel (perpendicular) to the magnetic field. Figure 4B reports the variation of S₁ with the shape parameter r assuming that the main axis of the oriented system is along z for ellipsoids and either along or perpendicular to z for bicelles.

A relative order parameter can also be defined, namely the ratio of the spectral widths of an experimental spectrum to that of a reference spectrum,

(32)

To validate the spectral simulations and the use of the order parameters presented in the previous section, several partially oriented lipid systems have been investigated. More specifically, we have first investigated the partial orientation of pure DMPC multilamellar vesicles as a function of temperature and of hydration level. This system is known to be partially deformed in high magnetic fields, and the results of the present study will show that the partial orientation can be quantitatively defined by the order parameter S₁. A similar approach will then be applied to DMPC:DHPC bicelles, a lipid system known to orient with the lipids oriented at 90° relative to the magnetic field. Finally, we will investigate two cases of partial parallel lipid orientation, namely the parallel orientation of DMPC:DHPC membranes in the presence of GA and the partial orientation of CL membranes in the presence of ADM.

Figure 5A presents the ³¹P NMR spectra of DMPC recorded at different conditions of concentration and temperature to study their partial orientation in the magnetic field. These spectra are characteristic of phospholipids in the liquid-crystalline phase. At a concentration of 40%, the spectrum is very close to the spectrum of a spherical distribution. However, lowering the concentration results in a decrease of the intensity of the lipids oriented at 0° relative to the magnetic field, which confirms the presence of partial orientation. It is also possible to note the appearance of an isotropic peak, which could be related to the formation of smaller structures in which the tumbling and the lateral diffusion correlation times become on the same order as the NMR acquisition time. Even if hydration seems to have the most important effect on the partial orientation, temperature also seems to induce a change of the orientation distribution.

Display large version of this figure

Figure 5

(A) Experimental ³¹P NMR spectra. (B) r Shape parameter for DMPC as a function of temperature and concentration (wt/wt %).

These spectral deformations were investigated by an orientation calculation. Information about the orientation in this system could be obtained only if the isotropic and anisotropic chemical shifts are evaluated precisely. The isotropic chemical shifts were obtained from MAS spectra but are not reported here, and the chemical shift anisotropy was evaluated directly on the static spectra. The CSA can be obtained in different ways, such as evaluating the second spectral moment. However, simulations have proven that an extremely accurate measurement of the CSA is not necessary for the orientational calculations (results not shown) and we have therefore evaluated this parameter by taking the chemical shift at 90% of the maximum spectral intensity. Our results indicate that both the CSA and the isotropic chemical shift vary linearly with temperature, with a transition at the gel to liquid-crystalline phase transition (results not shown). Then, an ellipsoidal parameter, r, was evaluted from the first spectral moment using Eq. (29A) and plotted as a function of the phospholipid concentration for two temperatures in Figure 5B. The ratio of the semi-axes c/a increases with temperature and DMPC concentration, indicating a higher orientation at high temperature and low concentration. This is related to the membrane elasticity, which increases with an increasing temperature and decreasing concentration.

The second system investigated in the present study is made of DMPC and DHPC at a molar ratio of 2.7:1. This system is supposed to form discoidal structures that orient their main axis at 90° relative to the magnetic field. Figure 6A shows the spectra obtained as a function of temperature. Below 25°C, the spectrum is composed of an isotropic peak that can be attributed to DHPC and of a broad spectrum that can be attributed to DMPC (Sanders and Schwonek, 1992). The intensities of the two subspectra correspond approximately to the molar fraction of DMPC and DHPC used in the sample preparation. From 25° to 35°C, the spectra indicate that the system becomes more and more oriented, i.e., the intensity corresponding to a perpendicular orientation of the phospholipids increases greatly. A second smaller peak, characterized by a different chemical shift, is also present and its chemical shift decreases with temperature.

Display large version of this figure

Figure 6

(A) Experimental ³¹P NMR spectra. (B) S₁ and S₂ order parameters and (C) S₁^t order parameter and t for DMPC:DHPC (2.7:1 molar ratio) as a function of temperature.

In a previous section, we have demonstrated that, as some authors have suggested (Sanders and Schwonek, 1992,Vold and Prosser, 1996), the smaller peak is due to phospholipids in the edges of the bicelle, most likely the short-chain lipid DHPC. However, the relative intensity of the two peaks changes with time, which indicates that these two peaks might not be solely attributed to DMPC and DHPC. At temperatures above 40°C, the smaller peak disappears at the expense of an isotropic peak. This corresponds to a well-known property of bicelles, namely the existence of a magnetic phase transition at 40°C (Sanders and Schwonek, 1992). At temperatures above 40°C, the proportion of phospholipids with a 90° orientation seems to increase with temperature and the intensity of the isotropic peak decreases.

The order parameters S₁ and S₂ have been calculated for these spectra using pure DMPC multilamellar vesicles as the reference system. More specifically, S₁ is calculated from the first spectral moment. In a previous section, we have demonstrated that, even in the presence of two distinct peaks, S₁ can be calculated if the peaks have the same isotropic chemical shift. This seems very likely because DMPC and DHPC have the same headgroup. In addition, even if the two peaks are not solely attributed to DMPC or DHPC, no fine structure is observed in the two peaks, again suggesting similar isotropic chemical shifts and CSA for the two lipids. Figure 6B shows the variation of S₁ as a function of temperature. S₁ has a value of 0 at 20°C, representing no orientation and, as the temperature increases, S₁ decreases to −0.45 at 35°C, representing an almost complete perpendicular orientation of the lipids. Then, S₁ increases to −0.35 at 50°C and decreases to −0.40 at 65°C. The greatest orientation is therefore at 35°C with a transition between 40 and 50°C. These results indicate that bicelles are well oriented between 30 and 40°C. This temperature range is similar to that obtained by other groups (Sanders and Schwonek, 1992,Losonczi and Prestegard, 1998,Ottiger and Bax, 1998) that demonstrated that bicelles are well oriented between 30 and 40–45°C. The slight difference observed in the upper limit temperature might be due to a slight sampling heating effect resulting from the high proton decoupling power used in the present study.

The results presented above indicate that S₁ is a good tool for representing the magnetic orientation of phospholipids because it is directly related to the shape of the membrane. Another order parameter that can be measured on these spectra is S₂, which is related to the dynamics in the system relative to a reference system. In this case, the reference system is pure DMPC. S₂ is plotted as a function of temperature in Figure 6B. In the supposed temperature range of existence of the bicelles (<40°C), the order parameter indicates that the system is less ordered than is pure DMPC. This is in agreement with the formation of small structures such as bicelles. At high temperatures, S₂ is equal to 1, indicating that the averaging of the chemical shift tensor is similar to that obtained for pure DMPC. The spectra are also broader and S₁ higher, showing that the system is most likely constituted of bigger structures less oriented than the bicelles at low temperature. These comments are based on the assumption that the S_tilt of the bicelles equals −0.5. In this system, the parameter S₂ therefore provides important and complementary information about the dynamics in the system.

To go further in the analysis of the system, we have used the proposed shape of the bicelles discussed in the Theory section to obtain the r (a/c ratio) value from the relative proportion of the two peaks at 35°C. Using Eq. (22), we found a value of 0.07, indicating that the semi-axis c is 15 times the value of the semi-axis a. Assuming that the bicelle thickness is 50Å, the total diameter of the bicelle is ∼750Å. In contrast, it is possible to determine the r shape parameter from the total S₁ value at 35°C. Using Eqs. (28), we found that r=0.05, indicating that the semi-axis c is 20 times the value of the semi-axis a, which gives a bicelle diameter of ∼1000Å. Therefore, these two different methods give approximately the same diameter, which corresponds to the proposed diameter for discoidal structures in lyotropic liquid-crystals (Forrest and Reeves, 1981). Another way to determine the r shape parameter is from the S₁ of the edge phospholipids. This value is plotted as a function of temperature in Figure 6C. The first remarkable feature is the linear relationship between S₁ and temperature between 25° and 45°C. However, it is surprising that all the S₁ values are below −0.125, the limit value. A way to explain this phenomenon is to suppose an exchange between the phospholipids in the torus and in the planes of the bicelles. The resulting order parameter could be represented by

(33)

One potential limitation in using the magnetically oriented bilayers described in the two previous sections for structural studies is that, because this system is characterized by a negative orientational order parameter (S₁≈−0.5), a well-resolved NMR spectrum with sharp lines will only be obtained if the molecule of interest undergoes fast axially symmetric motions. Otherwise, the NMR spectra will exhibit cylindrical powder patterns (Prosser et al). For this reason, the alignment of phospholipid bilayers with their director parallel to the magnetic field (i.e., with S₁=1) has been the goal of several research efforts. Recently, it was observed that the addition of small amounts of paramagnetic ions such as Eu³⁺ or Yb³⁺ to DMPC/DHPC bicelles results in systems in which the director is oriented parallel to the magnetic field (Prosser et al). Another way to obtain a parallel lipid orientation without these paramagnetic effects would be the addition of a molecule (protein, peptide, or drug) with a large positive Δχ (Sanders et al). This avenue was used in the present study. More specifically, we have first prepared bicelles in the presence of the transmembrane peptide GA due to its high content in aromatic residues and to its helical conformation. Figure 7A shows the spectra of DMPC:DHPC:GA as a function of temperature and time. These spectra clearly show the appearance of a spectral component at δ_‖ both with increasing temperature but also as a function of the time spent in the magnet.

Display large version of this figure

Figure 7

(A) Experimental ³¹P NMR spectra and (B) S₁ and r for DMPC:DHPC:GA (10:4.3:1 molar ratio) as a function of temperature and time. The spectra have been recorded every four hours at the following temperatures: (a) 25°C, (b) 30°C, (c) 35°C, (d) 40°C, (e) 45°C, (f to l) 50°C.

The comparison between the spectra presented in Figure 7A and the simulated spectra presented in Figure 2 and Figure 3 indicates that the shape of the partially oriented system is closer to a static parallel bicelle than to a parallel ellipsoid. To obtain more quantitative information about this orientation phenomenon, S₁ and S₂ were calculated using pure DMPC multilamellar vesicles as the reference system. The S₂ value is constant at 0.7 for all spectra. Thus, it can be considered that these systems are less ordered than pure DMPC. This is in agreement with the disordering effect also observed in ³¹P NMR spectra of unoriented DMPC lamellar vesicles in the presence of GA (Bélanger, A. and Auger, M., unpublished results). Figure 7B presents the S₁ calculation that provides quantitative information about the orientation of the lipid systems. Hence, S₁ goes up to 0.55 in the last spectrum, which is representative of a high orientational order. In addition, Figure 7B shows the r (a/c) ratio obtained from the order parameter S₁. For the highly oriented system at the end of the experiment, the calculation of the diameter from the r value indicates that the bicelle has a diameter four times bigger than its thickness. This observation is in agreement with the proposed size of bicelles (Vold and Prosser, 1996).

Another oriented system with the lipids parallel to the magnetic field is the CL:ADM complex at a 2:1 molar ratio. ADM is a highly aromatic molecule used currently in chemotherapy as an antineoplastic agent. This molecule is known to interact strongly with negatively charged lipids such as CL, a phospholipid found in the negatively charged cardiac cellular membranes. ADM has been shown to cause cardiac arrest if taken at high dosage. A model of interaction for this system has been proposed in the literature in which the ADM molecules are stacked on the membrane at an angle of 39° relative to the bilayer normal (Goormaghtigh et al). This arrangement is characterized by a large value of Δχ which favors its magnetic orientation. Figure 8A shows the spectra of the CL:ADM complex at a 2:1 molar ratio as a function of temperature. Significant magnetic orientation occurs in this system as the temperature increases, which is in agreement with the model of interaction proposed for this complex. Comparing these lineshapes with those presented in Figure 2 and Figure 3 indicates that these spectra could again be associated to a bicellar organization. However, the resolution is not as good and the spectrum is broader, characteristic of a distribution of CSA. Another satisfactory model could be an intermediate system composed of spherical vesicles with flat sections where the phospholipids are oriented parallel to the magnetic field.

Display large version of this figure

Figure 8

(A) Experimental ³¹P NMR spectra and (B) S₁ as a function of temperature for the CL:ADM (2:1 molar ratio) system. A line broadening of 150Hz was applied to these spectra.

Figure 8B shows the results of the order parameter calculations for the ADM:CL system. The value of S₁ increases gradually with temperature. However, the orientation does not become as high as that observed in the DMPC:DHPC:GA system. Using pure CL multilamellar vesicles as the reference system, the value of S₂ is ∼1 at all temperatures, indicating that the dynamics in this system are close to those of the pure CL system. These observations are in agreement with a broader distribution of orientations that is probably not associated to a perfect bicellar system. The r shape parameter has not been calculated because it is impossible to assume a perfect bicellar shape for this system.

The concept of magnetic orientation is not really new. However, a lack of methods to evaluate this phenomenon is evident in this area of research. Some authors (Forrest and Reeves, 1981) proposed order parameters to evaluate the extent of magnetic orientation, but these parameters are not really appropriate. More specifically, the measurement of the orientation via an order parameter such as S₂, as described above, is interesting but not complete. This order parameter could be indicative if a tilt of the bicelles or ellipsoids relative to the magnetic field axis is assumed. However, in general, it is more appropriate to suppose that the main axis of the deformed membrane is along or perpendicular to the magnetic field. In these cases, such order parameters will provide interesting information about the dynamics in the system. Therefore, the definition of another order parameter is really important. In the previous section, we proposed a new order parameter, S₁, derived from the measurement of the first spectral moment.

Spectral moments are well defined in the literature and are very easy to measure (Abragam, 1961). Therefore, an orientation parameter derived from this type of measurement is suitable. In fact, a change in the spectral lineshape due to partial orientation will affect the spectral moments. In principle, it could be possible to obtain orientation-dependent measurements from each of the spectral moments (M₁, M₂, M₃, M₄, …). However, the moments higher than the first one are very dependent on the spectral line broadening and only M₁ is not affected by symmetrical broadenings, such as Lorentzian or Gaussian broadenings. Therefore, only the use of M₁ is possible, even if M₁ is dependent on both the CSA and the isotropic chemical shift. With an evaluation of δ on a well-defined static spectrum and of δ_iso from a MAS spectrum, a really precise evaluation of S₁ is possible.

Because of the increasing use of both parallel and perpendicular bicelles as model membrane systems, the characterization of their shape and size is important. A model has recently been proposed based on ²H NMR measurements and on the calculation of the relative areas of the planes and edges of the bicelles (Vold and Prosser, 1996). There are also extensive studies of the position of DHPC on the edges of the bicelles by ²H and ³¹P NMR spectroscopy (Sanders and Schwonek, 1992,Vold and Prosser, 1996). However, there are no simulated spectra of such bicelles in the literature, and therefore, an attempt to simulate both the perpendicular and parallel orientation of bicelles seems valuable.

The experimental spectra obtained for bicelles with their main axes oriented parallel to the magnetic field are in agreement with simulated spectra, suggesting the validity of this model. For perpendicular bicelles made of DMPC and DHPC, the presence of a second peak with a frequency different from δ_iso is a convincing proof of the validity of the bicelle model. In addition, both the relative intensity of the two peaks and the S₁ calculation give the same r shape parameter, which corresponds to the proposed diameter of the bicelles. Finally, the values of S₁ obtained for the phospholipids in the edges of the bicelles suggest that there is a phase separation between the lipids in the planes and the torus of the bicelles and that there is a fast exchange process between them. These results from ³¹P NMR are in agreement with those obtained from ²H NMR measurements (Vold and Prosser, 1996).

A new order parameter, S₁, has been proposed in the present study to obtain quantitative information about the orientation of phospholipidic systems. This new tool can be helpful both to investigate the shape of lipid membranes and to study the effect of several parameters (temperature, hydration, addition of proteins or drugs, etc.) on the orientation of phospholipidic systems. A second order parameter, S₂, can provide complementary information about the dynamics in an oriented system. ³¹P NMR spectra have been simulated for both ellipsoid and bicellar systems in which the lipids are oriented either parallel or perpendicular to the external magnetic field. In addition, the S₁ and S₂ order parameters have been determined from the experimental spectra obtained for several systems.

More specifically, the general influence of temperature and concentration on orientation has been clearly demonstrated for pure DMPC multilamellar vesicles. In addition, the orientation and shape of bicelles made of DMPC:DHPC, in which the lipids are oriented perpendicular to the magnetic field, has been determined as a function of temperature. We have also shown in this study that the addition of GA to the DMPC:DHPC system induces an orientation of the lipids parallel to the magnetic field that seems to favor the formation of bicelles. Adriamycin also induces an orientation of the lipids at 0° relative to the magnetic field in CL bilayers, even in the absence of the short chain lipid DHPC. This indicates that a partial positive ordering can be obtained in lipid systems without the addition of paramagnetic ions and can be well characterized by the order parameter S₁ derived from the first spectral moment.