pubs.acs.org/Langmuir © 2009 American Chemical Society
)
Novel Type of Bicellar Disks from a Mixture of DMPC and DMPE-DTPA with Complexed Lanthanides Paul Beck,† Marianne Liebi,*,† Joachim Kohlbrecher,‡ Takashi Ishikawa,§ Heinz R€uegger, Peter Fischer,† Peter Walde,^ and Erich Windhab†
,#
† Laboratory of Food Process Engineering, ETH Zurich, Schmelzbergstrasse 9, 8092 Zurich, Switzerland, Laboratory for Neutron Scattering, ETH Zurich & Paul Scherrer Institute, 5232 Villigen PSI, Switzerland, § Department of Biology, ETH Zurich, Schafmattstrasse 20, 8093 Zurich, Switzerland, Laboratory for Inorganic Chemistry, ETH Zurich, Wolfgang-Pauli-Strasse 10, 8093 Zurich, Switzerland, and ^Department of Materials, ETH Zurich, Wolfgang-Pauli-Strasse 10, 8093 Zurich, Switzerland. # Heinz R€ uegger deceased June 28, 2009 )
‡
Received October 8, 2009 We report on the formation of bicelles from a mixture of dimyristoylphosphatidylcholine (DMPC) and the chelatorlipid dimyristoylphosphatidylethanolamine-diethylenetriaminepentaacetate (DMPE-DTPA) with complexed lanthanides, either thulium (Tm3þ) or lanthanum (La3þ). The two phospholipids used have the same acyl-chain length but differ in headgroup size and chemical structure. The total lipid concentration was 15 mM, and the molar ratio of DMPC to DMPE-DTPA was 4:1. The system was studied with small angle neutron scattering (SANS) in a magnetic field, cryotransmission electron microscopy (cryo-TEM), and 31P NMR spectroscopy. We found that, after appropriate preparation steps, that is, extrusion through a polycarbonate membrane followed by a cooling step, monodisperse small unilamellar disks (flat cylinders called bicelles) are formed. They have a radius of 20 nm and a bilayer thickness of about 4 nm and are stable in the investigated temperature range of 2.5-30 C. Fitting of SANS data with a form factor for partly aligned flat cylinders shows that the bicelles are slightly orientable in a magnetic field of 8 T if DMPE-DTPA is complexed with Tm3þ.
Introduction Bilayered micelles or “bicelles” are an important auxiliary system for the study of membrane protein structure with high resolution NMR. The term “bicelle” is used in this work to denote a small disklike aggregate, usually formed by a mixture of a long chain phospholipid and a detergent or short chain phospholipid.1 The most common bicelle formulation is DMPC/DHPC, where DMPC (1,2-dimyristoyl-sn-glycero-3-phosphocholine) forms the flat part of the disk and DHPC (1,2-dihexanoyl-sn-glycero-3phosphocholine) covers the edge.1 In addition to bilayered disks, mixtures of DMPC and DHPC can form other morphologies, depending on temperature, molar ratio of DMPC:DHPC, and lipid concentration.1-8 The origin of the different morphologies lies primarily in the temperature and concentration dependence of the miscibility of the two lipids.7 As the temperature is raised, DHPC becomes increasingly more soluble in DMPC; in consequence, less DHPC is available to cover the edges of the bicelles. As a result, much larger structures are formed, such as chiral *To whom correspondence should be addressed. E-mail: marianne.
[email protected]. (1) Katsaras, J.; Harroun, T. A.; Pencer, J.; Nieh, M. P. Naturwissenschaften 2005, 92, 355–366. (2) Nieh, M. P.; Glinka, C. J.; Krueger, S.; Prosser, R. S.; Katsaras, J. Langmuir 2001, 17, 2629–2638. (3) Nieh, M. P.; Glinka, C. J.; Krueger, S.; Prosser, R. S.; Katsaras, J. Biophys. J. 2002, 82, 2487–2498. (4) van Dam, L.; Karlsson, G.; Edwards, K. Biochim. Biophys. Acta 2004, 1664, 241–256. (5) van Dam, L.; Karlsson, G.; Edwards, K. Langmuir 2006, 22, 3280–3285. (6) Harroun, T. A.; Koslowsky, M.; Nieh, M. P.; de Lannoy, C. F.; Raghunathan, V. A.; Katsaras, J. Langmuir 2005, 21, 5356–5361. (7) Triba, M. N.; Warschawski, D. E.; Devaux, P. F. Biophys. J. 2005, 88, 1887– 1901. (8) Triba, M. N.; Devaux, P. F.; Warschawski, D. E. Biophys. J. 2006, 91, 1357– 1367.
5382 DOI: 10.1021/la903806a
nematic ribbons or perforated lamellar sheets,1,9 these types of structures are suitable for aligning in magnetic fields. The force causing alignment of such systems in a magnetic field is the anisotropy of the diamagnetic susceptibility of the phospholipids, resulting in a preferred orientation with the long molecular axis perpendicular to the applied field.10-13 The orientation of a single molecule is opposed by the thermal fluctuation kBT. In a bilayer, where the molecules are aggregated mostly parallel to each other, the magnetic anisotropy becomes additive and macroscopic orientation of the aggregates in magnetic fields is feasible.11 Such partially aligned phases provide anisotropic environments for solute materials, such as globular proteins, which slightly align. The determination of their threedimensional structures by NMR techniques profits under these conditions from the additional observation of scaled down dipolar coupling constants. Their magnitudes are sensitive to the lengths and the orientation, with respect to the magnetic field, of the internuclear vector between two NMR active isotopes. This information can be used to increase the accuracy of NMR structure determinations.14 The magnetic orientability of bicelles can be influenced by doping the membrane with paramagnetic lanthanides possessing a large magnetic moment.3,10,15 As such, the lanthanides are often (9) Soong, R.; Macdonald, P. M. Langmuir 2008, 24, 518–527. (10) Binnemans, K.; Gorller-Walrand, C. Chem. Rev. 2002, 102, 2303–2345. (11) Qiu, X. X.; Mirau, P. A.; Pidgeon, C. Biochim. Biophys. Acta 1993, 1147, 59–72. (12) Scholz, F.; Boroske, E.; Helfrich, W. Biophys. J. 1984, 45, 589–592. (13) Speyer, J. B.; Sripada, P. K.; Dasgupta, S. K.; Shipley, G. G.; Griffin, R. G. Biophys. J. 1987, 51, 687–691. (14) Bax, A.; Kontaxis, G.; Tjandra, N. Methods Enzymol. 2001, 339, 127–174. (15) Prosser, R. S.; Volkov, V. B.; Shiyanovskaya, I. V. Biophys. J. 1998, 75, 2163–2169.
Published on Web 11/23/2009
Langmuir 2010, 26(8), 5382–5387
Beck et al.
Article
anchored to the membrane through a complexing agent (DTPA, diethylenetriaminepentaacetate), which is covalently bound to a phospholipid headgroup (DMPE-DTPA). While DMPC/DHPC is the most widely studied bicellar system, related formulations have been investigated as well. It has been shown that bicelle formation can still be observed if the chain length of the two saturated lipids is varied to some extent.8,16-19 Furthermore, addition of small amounts of the monounsaturated lipid POPC (1-palmitoyl-2-oleoyl-sn-glycero3-phosphocholine) does not hinder the formation of bicellar structures.8,16 Moreover, it is possible to replace the short chain phospholipid DHPC by the bile salt analogue CHAPSO (3-(chloramidopropyl)dimethylammonio-2-hydroxy-1-propanesulfonate).20 By adding small amounts of charged lipids, such as DMPG, the long-term colloidal stability can be enhanced.15 The pH stability was increased by using ether-linked phospholipids.21-23 Formation of bicelles has also been found for mixtures of phosphatidylcholines and poly(ethylene glycol)-conjugated lipids (PEG-lipids).24-27 In such systems, the PEG-lipids are located preferably in the highly curved edge of the disk, whereas the phospholipids, often with cholesterol, form the planar part of the disk.25,26 In this Article, we report about an investigation on a mixture of DMPC and DMPE-DTPA with complexed lanthanides. The two phospholipids used here, DMPC and DMPE-DTPA, have the same acyl-chain length but differ in headgroup size. The system was studied with small angle neutron scattering (SANS) measurements, cryo-transmission electron microscopy (cryo-TEM), and 31 P NMR spectroscopy. SANS measurements were also performed in a magnetic field of 8 T at temperatures ranging from 2.5 to 30 C using DMPE-DTPA containing complexed thulium ions (Tm3þ). For the NMR measurements, Tm3þ was replaced by the diamagnetic lanthanum ion (La3þ) in order to prevent line shift and broadening.15
Materials and Methods Materials. The phospholipids, 1,2-dimyristoyl-sn-glycero-3phosphocholine (DMPC) and 1,2-dimyristoyl-sn-glycero-3-phosphoethanolamine-diethylenetriaminepentaacetate (DMPE-DTPA), were purchased as chloroform solutions from Avanti Polar Lipids (Alabaster, AL) and used without further purification. TmCl3 (99.9%), LaCl3 (99.9%), and D2O (99.9 atom % D) used for samples containing Tm3þ and D2O (99.9 atom % D, containing 0.05 wt % 3-(trimethylsilyl)propionic-2.2.3.3-d4 acid, sodium salt) used for samples containing La3þ were from Sigma-Aldrich (Buchs, Switzerland). Stock solutions of 10 mM TmCl3 in MeOH and of 10 mM LaCl3 in MeOH/H2O (19/1 v/v) were prepared. Bicelle Preparation. After weighing the lipid solution and lanthanides into a round-bottom flask, chloroform and methanol (16) De Angelis, A. A.; Opella, S. J. Nature Protocols 2007, 2, 2332–2338. (17) Hare, B. J.; Prestegard, J. H.; Engelman, D. M. Biophys. J. 1995, 69, 1891– 1896. (18) Wang, H.; Eberstadt, M.; Olejniczak, E. T.; Meadows, R. P.; Fesik, S. W. J. Biomol. NMR 1998, 12, 443–446. (19) Whiles, J. A.; Glover, K. J.; Vold, R. R.; Komives, E. A. J. Magn. Reson. 2002, 158, 149–156. (20) Sanders, C. R.; Prestegard, J. H. Biophys. J. 1990, 58, 447–460. (21) Aussenac, F.; Lavigne, B.; Dufourc, E. J. Langmuir 2005, 21, 7129–7135. (22) Cavagnero, S.; Dyson, H. J.; Wright, P. E. J. Biomol. NMR 1999, 13, 387– 391. (23) Ottiger, M.; Bax, A. J. Biomol. NMR 1999, 13, 187–191. (24) Johansson, E.; Engvall, C.; Arfvidsson, M.; Lundahl, P.; Edwards, K. Biophys. Chem. 2005, 113, 183–192. (25) Johansson, E.; Lundquist, A.; Zuo, S. S.; Edwards, K. Biochim. Biophys. Acta 2007, 1768, 1518–1525. (26) Johnsson, M.; Edwards, K. Biophys. J. 2003, 85, 3839–3847. (27) Sandstrom, M. C.; Johansson, E.; Edwards, K. Biophys. Chem. 2008, 132, 97–103.
Langmuir 2010, 26(8), 5382–5387
were evaporated under a rotary evaporator followed by residual solvent removal under high vacuum overnight. In the following, the dry lipid film was hydrated with D2O and vortexed until a homogeneous suspension was obtained. Since La3þ could not be dispersed completely, the pH of the suspension was adjusted to pH = 7 with 0.1 M NaOH in H2O. Five repeated freezethaw cycles were carried out by plunging the flask into liquid nitrogen followed by slow heating above the phase transition temperature (about 40-45 C). The extrusion was performed under nitrogen at 40 C with “The Extruder” from Lipex Biomembranes (Vancouver, Canada) and Nucleopore polycarbonate membranes from Sterico (Dietikon, Switzerland).28,29 The suspension was first passed 10 times through a membrane with 200 nm pores, followed by 10 times through a membrane with 100 nm pores. All bicellar samples were prepared with a DMPC: DMPE-DTPA:lanthanide molar ratio of 4:1:1 and a total lipid concentration of 15 mM (denoted as DMPC/DMPE-DTPA 3 Tm or DMPC/DMPE-DTPA 3 La). Samples were stored before measurements at room temperature and were stable over at least 1 week. Cryo-TEM. The samples were analyzed by cryo-transmission electron microscopy (cryo-TEM) following the procedure described previously.30 A volume of 3.5 μL of the bicellar suspension was mounted onto holey carbon grids (Quantifoil, Jena, Germany), blotted to make thin aqueous films under controlled temperature and humidity, followed by plunging into liquid ethane at the temperature of liquid nitrogen. For samples frozen from room temperature and from 30 C, a VitrobotTM apparatus (FEI, Eindhoven, The Netherlands) was used. For the samples frozen from 5 C, the sample was precooled in a water bath and the manual blotting apparatus was transferred into a cold room with 5 ( 0.5 C. Every part of the equipment coming into contact with the sample was held for at least 4 h in the cold room prior to sample freezing. The grids were examined at the temperature of liquid nitrogen using a cryo-holder (model 626, Gatan) and a Tecnai G2 F20 microscope (FEI) equipped with a field emission gun and Tridem energy filter (Gatan) operated at an accelerating voltage of 200 kV. The data were recorded by using a 2048 2048 CCD camera (Gatan). SANS. SANS experiments were performed on the SANS-I beamline at PSI, Villigen, Switzerland using a super conductive magnet with horizontal field of 8 T perpendicular to the neutron beam. The neutron wavelength was fixed at 8 A˚. Data were collected on a two-dimensional 3He detector at distances of 2, 6, and 18 m covering a momentum transfer of 0.003 e q e 0.15 A˚-1. After correction for background radiation, empty cell scattering, and detector efficiency, the 2D intensity maps were radial averaged and sectoral averaged with an opening angle of 15 perpendicular and parallel to the magnetic field direction. Data were fitted with a form factor for partly aligned flat cylinders. SANS Model for Partly Aligned Flat Cylinders. The scattering intensity of a flat cylinder Icyl(q) with a given orientation γ relative to the scattering vector is given by 0
12 qL sin cos γ 2 J1 ðqR sin γÞ B C Icyl, γ ðq, γÞ ¼ @2πR2 LΔη A qL qR sin γ cos γ 2
ð1Þ
where γ is the angle between the scattering vector q and the cylinder axis n.31 L is the length of the cylinder, R is its radius, Δη is the scattering length density contrast relative to the solvent, and J1(x) is the first order Bessel function of the first kind. γ can be calculated (see Figure 1) from the orientation (θ,φ) of the cylinder and the direction of the scattering vector ψ in (28) (29) (30) (31)
Mui, B.; Chow, L.; Hope, M. J. Methods Enzymol. 2003, 367, 3–14. Walde, P.; Ichikawa, S. Biomol. Eng. 2001, 18, 143–177. Ishikawa, T.; Sakakibara, H.; Oiwa, K. J. Mol. Biol. 2007, 368, 1249–1258. Hayter, J. B.; Penfold, J. J. Phys. Chem. 1984, 88, 4589–4593.
DOI: 10.1021/la903806a
5383
Article
Beck et al.
Figure 1. Sketch which defines the angles of the orientation n of partly aligned disks relative to the scattering vector q. the plane of the detector by 1 0 1 0 cos ψ cos θ q3n @ A @ A cos γ ¼ ¼ 3 sin θ sin φ 0 jqjjnj sin θ cos φ sin ψ ¼ cos ψ cos θ þ sin ψ sin θ cos φ
ð2Þ
If the orientation distribution of the orientation vector n is described by p(θ,φ), the scattering intensity is given by Z π Z 2π Icyl ðqÞ ¼ dθ dφ sinðθÞ pðθ, φÞ 0
0
B @2πR2 LΔη
0
12
J1 ðqR sin γÞ sin qR sin γ
qL 2 cos γ qL 2 cos γ
C A
ð3Þ
For this form factor, it is assumed that the orientation distribution is independent of φ, that is, p(θ,φ) = p(θ) and that p(θ) = p(π - θ), which means that turning the cylinder by 180 results in the same scattering intensity. Here, we are assuming a Maier-Saupe orientation probability distribution which reads as pffiffiffi K 2 pðθÞ ¼ K pffiffiffi eK cos 2e Dð KÞ
θ
ð4Þ
where D(x) is the Dawson R integral. The probability orientation distribution is normalized π0 p(θ) sin θ dθ = 1. The parameter κ is the only fit parameter of the orientation distribution and is here the ratio between the magnetic potential energy and the thermal energy K ¼
-Emag kB T
ð5Þ
where kB is the Boltzmann constant (kB = 1.306 10-16 erg K-1 in CGS units). NMR. 31P NMR spectra were taken from a mixture of DMPC and DMPE-DTPA with complexed lanthanum (La3þ) in D2O with a total lipid concentration of 15 mM, and the pH was adjusted to pH = 7. After preparation, including rehydration, freeze-thaw cycles, and extrusion (see above), the samples were kept at room temperature. The spectra were acquired at 283 MHz using a Bruker Biospin spectrometer operating with a 16.4 T field. Temperature was adjusted with a flow of cold nitrogen gas, heated to the appropriate temperature. Samples were measured at 22, 5, 22, 40, and again 22 C. Additionally, a sample was prepared where the cooling cycle (22, 5, 22 C) was performed outside of a magnetic field in an external water bath.
Results and Discussion In the context of a study of possible effects of magnetic fields on phospholipid vesicles,32 we found that a mixture of DMPC and (32) Beck, P.; Liebi, M.; Kohlbrecher, J.; Ishikawa, T.; R€uegger, H.; Zepik, H.; Fischer, P.; Walde, P.; Windhab, E., J. Phys. Chem. Submitted for publication.
5384 DOI: 10.1021/la903806a
Figure 2. Cryo-TEM micrograph of the sample used for the SANS measurements showing bicelles formed by DMPC and DMPE-DTPA complexed with Tm3þ at 5 C. Arrows point to bicelles edge-on (a) and face-on (b). DMPC:DMPE-DTPA:Tm, 4:1:1, lipid concentration 15 mM. Scale bar = 200 nm.
DMPE-DTPA with complexed lanthanides in a molar lipid ratio of 4:1 and a lipid concentration of 15 mM in water forms nonvesicular structures. These structures were investigated using SANS, cryo-TEM, and NMR spectroscopy. SANS and cryoTEM are complementary methods for structure determination of self-assembly systems. While SANS yields average information over the whole sample volume, cryo-TEM is a model independent measurement method, whereby individual objects can be observed directly. Cryo-TEM Measurements. Cryo-TEM micrographs were taken from a mixture of DMPC and DMPE-DTPA with complexed thulium (Tm3þ), abbreviated as DMPC/DMPE-DTPA 3 Tm. Figure 2 shows small disks of phospholipids (flat cylinders called bicelles) frozen from 5 C, which are randomly oriented. Some of the bicelles are seen edge-on (a in Figure 2) or face-on (b in Figure 2). The bicelles have a radius of roughly 20 nm. CryoTEM micrographs of DMPC/DMPE-DTPA 3 La frozen from 22 C are shown in Figure 3. The presence of bicelles is again obvious with bicelles seen edge-on or face-on. Figure 3A was taken at a tilting angle of 0, whereas the same sample spot in Figure 3B was tilted at 30 (see the Supporting Information for a detailed explanation). The ice crystal (arrow b) in both micrographs of Figure 3 can be taken as a reference point; the bicelle next to it (arrow a) is seen edge-on in Figure 3A and face-on in Figure 3B. Micrographs taken from DMPC/DMPE-DTPA 3 La frozen from 30 C showed bicelles corresponding in shape and size to those shown in Figure 3 (data not shown). The importance of the extrusion step during preparation is illustrated in Figure 4. Figure 4A is an electron micrograph of a sample containing DMPC and DMPE-DTPA complexed with Tm3þ after rehydration and before freeze-thaw cycles. Figure 4B shows an electron micrograph of the same sample after five freeze-thaw cycles, before extrusion. Predominantly large lamellar structures are seen in both micrographs, and no bicelles were found. The small structures in Figure 4A could be either ice contaminations from the freezing procedure or phospholipid micelles, which would be destroyed in the following freeze-thaw cycles. Langmuir 2010, 26(8), 5382–5387
Beck et al.
Article
Figure 3. Cryo-TEM micrographs of bicelles formed by DMPC and DMPE-DTPA complexed with La3þ at 22 C. Image A was taken at 0 tilting angle, and image B was taken at a tilting angle of 30 around the axis shown by the solid line. Arrow a points to the same bicelle edge-on (A) and face-on (B). Arrow b points to an ice crystal as reference point. DMPC:DMPE-DTPA:La, 4:1:1, lipid concentration 15 mM. Scale bar = 100 nm.
Figure 4. Cryo-TEM micrograph of DMPC/DMPE-DTPA 3 Tm at 22 C. Image A was taken from a sample before freeze-thaw cycles, and image B after five freeze-thaw cycles but before the extrusion step. No bicelles were found. DMPC:DMPE-DTPA:Tm, 4:1:1, lipid concentration 15 mM. Scale bar = 200 nm.
SANS Measurements. SANS measurements were carried out with the same repeatedly frozen-thawed and extruded sample of DMPC/DMPE-DTPA 3 Tm as used for cryo-TEM. The experiments were carried out in the presence of a magnetic field of 8 T. After an initial measurement at 30 C, the temperature was lowered stepwise to 2.5 C. Measurements were made at 30, 22, 17.5, 15, 10, 7.5, 5, and 2.5 C. The SANS scattering curves were fitted with a model for partly aligned flat cylinders, in accordance to the cryo-TEM micrographs showing small disks (bicelles) and a slightly anisotropic scattering pattern, indicating an orientation in the magnetic field. Figure 5 shows the SANS sectoral intensity average of DMPC/DMPE-DTPA 3 Tm at 2.5 C and 8 T from overlaid curves from 18, 6, and 2 m detector distance, as well as the corresponding fit (see the Supporting Information for corresponding 2D scattering pattern). The bicelle radius R for the best fit was 17 nm and the bilayer thickness L was 4.3 nm, in good agreement with the size of the bicelles found by cryo-TEM. The parameter κ from the orientation distribution (see eqs 4 and 5) was fitted to be 0.42. The corresponding orientation distribution (inset Figure 5) indicates that the probability of bicelles oriented with the normal parallel to the magnetic field (θ = 0) is slightly higher than the probability of an orientation perpendicular to the magnetic field (θ = π/2). SANS measurements at temperatures from 5 to 22 were only made at 18 m detector distance. While this q-range contains information about the overall size of the bicelles, no information Langmuir 2010, 26(8), 5382–5387
Figure 5. Sectoral averaged SANS curves of DMPC/DMPEDTPA 3 Tm at 2.5 C and corresponding fit with a model for partly aligned flat cylinders. Filled symbols mark the 15 horizontal sector in direction of the magnetic field, and open symbols the 15 vertical sector perpendicular to the magnetic field. Fit results indicate an orientation with the bilayer normal parallel to the magnetic field as shown in the orientation distribution (inset).
is given about the thickness of the bicelle and an exact fit of the orientation is not possible. Guinier fits at all temperatures showed no significant difference in the bicelle radius and no particular DOI: 10.1021/la903806a
5385
Article
Beck et al.
Figure 6. Radial averaged SANS curves of DMPE-DTPA 3 Tm and corresponding fit with a model for an ellipsoidal structure. Fit results indicate the presence of ellipsoidal micelles with semiprincipal axes of 2.2 nm and equatorial semiaxis of 3.3 nm. Inset shows the hard sphere structure factor. Lipid concentration was 10 mM, and measurements were carried out at 22 C and 0 T.
trend was observed. At 30 C, additional SANS measurements were carried out with a detectoral distance of 6 m. Like the curves obtained at 2.5 C, these resulting sectoral averaged SANS curves could be fitted to a model for partly aligned flat cylinders. The bicelle radius R for the best fit was 18.5 nm, the bilayer thickness L was 3.9 nm, and the parameter κ from the orientation distribution was 0.22. These results indicate that at 30 C bicelles are slightly bigger and less thick than those at 2.5 C, as can be expected for lipids in the fluid-analogue LR phase above the main phase transition temperature of DMPC Tm = 23.6 C.33 Faster lipid motion at higher temperatures may explain the less pronounced orientation. In addition, SANS measurements have been carried out with a sample consisting only of DMPE-DTPA with complexed Tm3þ in a molar ratio of 1:1 and a lipid concentration of 10 mM. The measurement was conducted at 22 C and 0 T. Figure 6 shows the full detector radial intensity average and the corresponding fit with an ellipsoidal form factor, including a hard sphere structure factor (shown in the inset).34 The semiprinicipal axis of the ellipsoid for the best fit was 2.2 nm, and the equatorial semiaxis was 3.3 nm. This result indicates that pure chelator-lipid with complexed lanthanide in aqueous solution forms micelles. The micelles are not perfectly spherical but are slightly deformed, forming an oblate ellipsoid. 31 P NMR Measurements. 31P NMR spectra were taken from the sample analyzed by cryo-TEM shown in Figure 3 (DMPC/DMPE-DTPA 3 La). Lanthanum instead of thulium was used in order to prevent shifts and broadenings of the NMR line. The spectra shown in Figure 7 were recorded in sequence from top to bottom. All spectra showed a single broad resonance at the position of the isotropic chemical shift. This is in agreement with the presence of bicelles and in contrast to what would be expected for oriented structures or larger entities such as vesicles. The first measurement (Figure 7A), taken right after extrusion at room temperature, shows an isotropic peak with a width of 2.0 kHz. Cooling the sample to 5 C increased the width to 3.5 kHz (Figure 7B), typical for slower phospholipid motion at this temperature. Reheating the sample (Figure 7C) resulted in a spectrum differing from the one recorded before the cooling step (Figure 7A). The peak is now considerably narrower (1.2 kHz), (33) Koynova, R.; Caffrey, M. Biochim. Biophys. Acta 1998, 1376, 91–145. (34) Aswal, V. K.; Kohlbrecher, J. Chem. Phys. Lett. 2006, 424, 91–96.
5386 DOI: 10.1021/la903806a
Figure 7.
31
P NMR spectra of DMPC/DMPE-DTPA 3 La bicelles at (A) T = 22 C, right after extrusion step, (B) 5 C, (C) 22 C, (D) 40 C, and (E) 22 C. The difference in peak width between (A) and (D) indicates the irreversible decrease of the bicelle size. Heating of the sample up to 40 C causes no irreversible change in the 22 C spectra (compare C and E). The isotropic peak at -1 ppm corresponds to small and fast tumbling bicelles.
indicating that an irreversible change in the size of the bicelles occurred. If the cooling cycle is performed outside of a magnetic field in a water bath (10 h at 5 C), an identical NMR spectrum as the one shown in Figure 7C is obtained. This size reduction of the bicelles by a cooling step is in line with the size evaluation of cryoTEM micrographs. The bicelle radius in the micrograph of a sample before the cooling step was 20 ( 8 nm (n = 415), whereas bicelles in the micrographs of the sample after the cooling step had a size of 18 ( 5 nm (n = 1136). Heating the sample to 40 C (Figure 7D) and cooling it back to room temperature (Figure 7E) did not lead to any further changes in the NMR spectrum (compare Figure 7C and E). All NMR spectra recorded are in agreement with small bicelles, whose rates of tumbling determine the width of the resonance. Cooling the sample to 5 C for around 10 h led to a homogenization and a slight size reduction of the bicelles. Structural information obtained with all three methods used, SANS, cryo-TEM, and 31P NMR, is consistent with small dispersed disks in an aqueous mixture of DMPC/DMPE-DTPA (4:1) complexed with either Tm3þ or La3þat 15 mM total lipid concentration. Important for the preparation is an extrusion step at 40 C. Furthermore, a cooling step at 5 C can help to homogenize the disks to an average diameter of 36 nm. Samples with thulium can be oriented in a magnetic field (8 T), as shown with SANS, whereas samples with lanthanum show no orientation in the magnetic field as shown with NMR. Figure 8 shows a sketch of a cross section of such a bicelle (edge-on view). We assume that the plane of the bicelles consists mainly of DMPC, whereas the necessary local curvature to cover the edge of the bilayer disk is most likely induced by the large headgroup of DMPE-DTPA complexed with either Tm3þ or La3þ. Langmuir 2010, 26(8), 5382–5387
Beck et al.
Article
Figure 8. Sketch of a bicelle formed by a mixture of DMPC, aggregated preferentially in the bilayered center of the disk and DMPE-DTPA complexed with a lanthanide covering mainly the edge (cross section of an edge-on view).
The driving force for the lipid segregation is the different packing parameters of the two lipids due to the large headgroup of DMPE-DTPA 3 lanthanide as compared to DMPC. SANS measurements of a sample with only DMPE-DTPA 3 lanthanide showed the formation of ellipsoidal micelles, indicating the preference of this lipid to assemble in highly curved assemblies, as in the edge region of the bicelle. This finding is in line with Johnsson and Edwards,26 who described for a lipid mixture containing poly(ethylene glycol) (PEG)-lipids such a partial component segregation, leading to the formation of disklike micelles, with PEG-lipids covering the highly curved rim. Recently, this lipid segregation was confirmed by SANS using a contrast matching technique with deuterated phospholipids.35 The magnetic orientation energy of aggregates in a magnetic field H can be calculated as follows 1 Emag ¼ - μ0 nΔχH 2 2
ð6Þ
where n is the aggregation number, Δχ is the molecular diamagnetic susceptibility anisotropy of individual phospholipid molecules, and μ0 is the magnetic constant (μ0 = 1 in CGS units).36 With eq 5 and the fit parameter of the orientation distribution κ = 0.42, the magnetic orientation energy Emag of the bicelles described here at 2.5 C is calculated to be -1.60 10-14 erg (CGS). Some values of the diamagnetic susceptibility anisotropy of DMPC can be found in the literature. Scholz et al.12 measured the magnetic polarizability per unit area of bilayer bΔχ in the LR phase to be -0.58 10-14 cm (CGS), where b is the membrane thickness, which gives a value of the magnetic susceptibility anisotropy Δχ of -1.44 10-8 erg cm-3 G-2 (CGS) on the assumption of a membrane thickness of 4.02 10-7 cm in the LR phase.37 According to Prosser et al.,15 Δχ of DMPE-DTPA 3 Tm is expected to be at least 155 times that of DMPC and of opposite sign. The molecular magnetic susceptibility anisotropy Δχ of DMPE-DTPA 3 Tm is therewith calculated to be at least 2.70 10-27 erg G-2 (CGS), under the assumption of a molecular volume of 1.2 10-21 cm3 in the LR phase (molecular volume of DMPC = 1.101 10-21 cm3 (ref 38) plus molecular volume of DTPA = 0.108 10-21 cm3 (ref 39)). With these values, the aggregate number of DMPE-DTPA 3 Tm needed for the magnetic (35) Lundquist, A.; Wessman, P.; Rennie, A. R.; Edwards, K. Biochim. Biophys. Acta 2008, 1778, 2210–2216. (36) Yamaguchi, M.; Tanimoto, Y. Magneto-Science. Magnetic Field Effects on Materials: Fundamentals and Applications; Springer: Berlin, Heidelberg, New York, 2006. (37) Pencer, J.; Nieh, M. P.; Harroun, T. A.; Krueger, S.; Adams, C.; Katsaras, J. Biochim. Biophys. Acta 2005, 1720, 84–91.
Langmuir 2010, 26(8), 5382–5387
orientation at 8 T found in SANS would be around 1850 molecules. This is about 2.5 times higher than the number of DMPE-DTPA 3 Tm molecules expected in a bicelle with a radius of 17 nm and 4.3 nm thickness and a molar ratio of DMPC:DMPE-DTPA 3 Tm of 4:1 in the bicelle. This deviation can be rationalized by the following three reasons: First, the values for the molecular volume and the magnetic susceptibility anisotropy of DMPE-DTPA 3 Tm are based on several assumptions, as described above. The true value may differ. Second the value of Δχ of phospholipids below the phase transition temperature Tm is expected to be higher than the provided values in the LR phase, caused by the higher molecular order in the solid-analogue state.12 This is in accordance with the less pronounced orientation measured at high temperatures. Finally, it is possible that the bicelles are partly aggregated, forming small bicellar stacks, which would enhance the magnetic orientability significantly, even though no such stacks have been seen in cryo-TEM.
Conclusion In this report, we show that bicelles can be formed from a mixture of DMPC and DMPE-DTPA 3 lanthanide. Conventional phospholipid bicelles consist of a long chain phospholipid plane with a short chain lipid, covering the edges.1 Here, we show that bicelles can be formed with only one type of fatty acid tail, myristoyl. We assume that partial lipid segregation occurs, due to the large headgroup of the DMPE-DTPA 3 lanthanide, which is located preferentially in the highly curved edge region of the bicelle. DMPC on the other hand is aggregated mainly in the planar center of the bicelle (see Figure 8). The presence of complexed cations is a prerequisite for bicelle formation, as supported by ongoing SANS study of samples without lanthanides, which shows rodlike structures instead of bicelles. Weak alignability in a magnetic field of 8 T is observed if the paramagnetic Tm3þ is used, while there is no indication of aligning with the diamagnetic La3þ. For the system described here, the method of preparation, including extrusion through a polycarbonate membrane followed by a cooling step, is essential to obtain bicelles. The so-formed bicelles are stable over a temperature range of 2.5-30 C for at least 1 week as shown with SANS and cryo-TEM. Acknowledgment. We acknowledge the Electron Microscopy Center of the ETH Zurich (EMEZ) for technical support. Supporting Information Available: Sketch illustrating how tilted cryo-TEM micrographs were taken; SANS 2D scattering pattern and corresponding sectoral intensity average including error bars of the discussed bicellar system. This material is available free of charge via the Internet at http:// pubs.acs.org. (38) Nagle, J. F.; Tristram-Nagle, S. Biochim. Biophys. Acta 2000, 1469, 159– 195. (39) D’Arceuil, H. E.; de Crespigny, A. J.; Pelc, L.; Howard, D.; Alley, M.; Seric, S.; Hashiguchi, Y.; Nakatani, A.; Moseley, M. E. Magn. Reson. Imaging 2004, 22, 1243–1248.
DOI: 10.1021/la903806a
5387
