Langmuir 2008, 24, 709-713
709
Growth Process and Molecular Packing of a Self-assembled Lipid Nanotube: Phase-Contrast Transmission Electron Microscopy and XRD Analyses Hiroharu Yui,*,†,‡ Hiroyuki Minamikawa,‡,§ Radostin Danev,| Kuniaki Nagayama,| Shoko Kamiya,‡ and Toshimi Shimizu*,‡,§ Department of Chemistry, Faculty of Science, Tokyo UniVersity of Science (TUS), Funagawara-machi 12, Ichigaya, Shinjuku-ku, Tokyo 162-0826, Japan, SORST, Japan Science and Technology Agency (JST), Tsukuba Central 5, 1-1-1 Higashi, Tsukuba, Ibaraki 305-8565, Japan, Nanoarchitectonics Research Center (NARC), National Institute of AdVanced Industrial Science and Technology (AIST), Tsukuba Central 5, 1-1-1 Higashi, Tsukuba, Ibaraki 305-8565, Japan, and Okazaki Institute for IntegratiVe Bioscience, National Institute of Natural Sciences, 5-1 Higashiyama, Myodaiji-cho, Okazaki, Aichi 444-8787, Japan ReceiVed August 12, 2007. In Final Form: October 16, 2007 Phase-contrast transmission electron microscopy (PC-TEM) and quick freezing method have been combined to study the initial growing process of a self-assembled lipid nanotube in water. The PC-TEM enabled us to detect thin lamellar edge structure and the very fast growth of the newborn edge to a thin tube with high contrast. The thin tube acts as a core structure for further growth into thick complete lipid nanotube. The initially formed nanotube structure is denoted as a “core tube”. The core tube has uniform wall structure that consists of five lamellar layers and the inner and outer diameters of the core tube are 130 and 180 nm, respectively. The evaluated lamellar spacing of 4.6 nm is well compatible with that measured by X-ray diffraction. We also discussed the molecular packing of the nanotube from the pitch angle determined by the PC-TEM images, X-ray diffraction pattern in wide-angle region, and IR spectroscopy. The subcell structure of the nanotube is assigned to an orthorhombic type. The twisting angle between the neighboring lipid molecules is determined as ca. 0.26° for the first time; it is a crucial parameter for the formation of a lipid nanotube in chiral packing but has not been elucidated before.
Introduction Lipid nanotubes (LNTs) have been of great interest because of their potent applications to one-dimensional fabrication of nanomaterials and mesoscale biocompatible host materials for relatively large biomolecules such as proteins and DNAs.1-3 For various applications, it is crucial to control the dimensions of the LNTs by appropriate designing of the structure of lipid molecules, namely, building blocks, and control the packing between them. To elucidate the relation between the molecular structure and the resultant morphology of their self-assembled structures, it is essential to observe the self-assembling process.4-8 Especially, the molecular packing and the morphology of the initial core structures are key factors for the determination of the final morphology of the self-assembled structure and its dimensions. However, the initial core structure of the LNTs, which is expected to have nanoscale dimensions and consist of only a few lamellar * To whom correspondence should be addressed. † Tokyo University of Science. ‡ Japan Science and Technology Agency. § National Institute of Advanced Industrial Science and Technology. | National Institute of Natural Sciences. (1) Shimizu, T.; Masuda, M.; Minamikawa, H. Chem. ReV. 2005, 105, 1401, and references cited therein. (2) Schnur, J. M. Science 1993, 262, 1669. (3) Fuhrhop, J. H.; Helfrich, W. Chem. ReV. 1993, 93, 1565. (4) Yamada, K.; Ihara, H.; Ide, T.; Fukumoto, T.; Hirayama, C. Chem. Lett. 1984, 1713. (5) Yager, P.; Schoen, P. E. Mol. Cryst. Liq. Cryst. 1984, 106, 371. (6) Nakashima, N.; Asakuma, S.; Kunitake, T. J. Am. Chem. Soc. 1985, 107, 509. (7) Thomas, B. N.; Lindemann, C. M.; Clark, N. A. Phys. ReV. E 1999, 59, 3040. (8) Yoshida, K.; Minamikawa, H.; Kamiya, S.; Shimizu, T.; Isoda, S. J. Nanosci. Nanotechnol. 2007, 7, 960.
sheets without metal molecules, gives no high-contrast transmission electron microscope images in general. Here, we combined one kind of phase-contrast TEM “Zernike Phase-Contrast TEM” (ZPC-TEM) with a quick freezing technique to probe the initial structures of a self-assembled LNT.9-11 The LNT was prepared by the self-assembly of the N-(11-cis-octadecenoyl)-β-D-glucopyranosylamine (1)12 and its initial edge structure and the core structure were observed by the ZPC-TEM. In addition, by considering ZPC-TEM images, XRD patterns, and IR spectrum, we discussed the molecular packing of the self-assembled LNT from 1. The twisting angle between the neighboring lipid molecules, which is a key parameter for the nanotube formation based on chiral self-assembly,13,14 was evaluated for the first time. Experimental Section Sample Preparation. N-(11-cis-Octadecenoyl)-β-D-glucopyranosylamine (1) (Figure 1a) was synthesized and characterized as described previously.12 We dispersed 5 mg of 1 in 100 mL of water at 100 °C for 12 h. Then the aqueous dispersion was allowed to cool to room temperature at the rate of 2.5 °C/min for the first 30 min. When the sample temperature reached 60 °C, an aqueous dispersion containing the LNTs was sampled and instantaneously transferred on a copper grid coated with holey carbon film. The sampling (9) Nagayama, K. AdV. Imagin Electron Phys. 2005, 138, 69. (10) Kaneko, Y.; Danev, R.; Nitta, K.; Nagayama, K. J. Electron Microsc. 2005, 54, 1. (11) Danev, R.; Okamura, H.; Usuda, N.; Kametani, K.; Nagayama, K. J. Biol. Phys. 2002, 28, 627. (12) Kamiya, S.; Minamikawa, H.; Jung, J. H.; Yang, B.; Masuda, M.; Shimizu, T. Langmuir 2005, 21, 743. (13) Selinger, J. V.; Spector, M. S.; Schnur, J. M. J. Phys. Chem. B 2001, 105, 7157. (14) Helfrich, W.; Prost, J. Phys. ReV. A 1998, 38, 3065.
10.1021/la702488u CCC: $40.75 © 2008 American Chemical Society Published on Web 01/08/2008
710 Langmuir, Vol. 24, No. 3, 2008
Figure 1. (a) Molecular structure of N-(11-cis-octadecenoyl)-βD-glucopyranosylamine 1. (b) A conventional TEM image of the self-assembled lipid nanotube as a final morphology from the lipid molecule 1. temperature (60 °C) is just below the gel-to-liquid-crystalline phase transition temperature (62 °C).12 An edge structure was observed only at the sampling temperature of 60 °C, along with a number of core nanotube structures. Below the temperature (e.g., 59 °C), we can observe only further grown nanotubes. After careful removal of excess liquid with the tip of filter paper, the sample was frozen rapidly in liquid ethane and was transferred into the ZPC-TEM microscope using a LEICA EM CPC cryo-station system (Leica Microsystems, Vienna, Austria). The remaining aqueous sample was further cooled down to room temperature, and after 3 days, the final morphology of the LNTs was measured with a conventional TEM microscope (JEOL JEM2000EXII, acceleration voltage: 200 kV, magnification: 5000×). The inner and outer diameters of the final morphology of the LNTs were determined by averaging 80 LNTs. Phase-Contrast TEM Measurements. The initial structures were observed by a transmission electron microscope equipped with a specially designed phase plate prepared from amorphous carbon films for the phase contrast, which corresponds to the phase contrast (PC) in light microscopy.11 The Zernike phase contrast (ZPC) can supply high-contrast TEM images without staining.9,10 As the PC does, the ZPC can display nanostructures of thin specimen objects in a topographical manner. The experiments were carried out on a JEOL JEM-3100FFC electron microscope operated at 300 kV acceleration voltage with the ZPC phase plate. The microscope was equipped with a field-emission gun and omega-type energy filter. The objective lens parameters were spherical aberration coefficient 5 mm and chromatic aberration coefficient 4.7 mm. All observations were performed with a magnification of 66000× at a CCD camera plane with a resolution of 0.45 nm/pixel and an electron dose of approximately 630 e-/nm2 in zero-loss filtering mode. The energy window width was set at 10 eV. A special aperture holder with heating was used to support the phase plates. To avoid contamination, the phase plates were kept at approximately 200 °C at all times. All images were recorded with a Gatan MegaScan795 2K × 2K changedcoupled device (CCD) camera.10 Digital Micrograph, supported by Gatan, was used for image analysis. XRD Measurement. A suspension of the LNTs was centrifuged to collect the nanotubes. The obtained wet LNTs were transferred into the bottom of a quartz capillary (Glas, 1.5 mm in outer diameter, 0.01 mm in wall thickness), together with a small drop of water. The capillary was flame-sealed at its upper end to avoid water evaporation. The X-ray diffraction experiment was performed on a Rigaku R-AXIS IV X-ray diffractometer, with Cu KR radiation (λ ) 0.1542 nm, 40 kV, 30 mA) monochromated with graded elliptic multilayer optics. The diffraction pattern was recorded with an imaging plate (Fuji-
Yui et al.
Figure 2. ZPC-TEM image corresponding to an initial structure of a core lipid nanotube. Scale bar indicates 100 nm in length. Photo Film) in a flat camera (camera length: 145.5 mm). The measurement was carried out at 25 °C under a 60 min exposure. ATR-FT-IR Measurement. IR spectrum was measured on an ATR diamond cell (PIKE Technologies, Inc.) at ambient temperature using an FT-IR spectroscope (FT/IR-680plus, JASCO Co., Japan).
Results and Discussion Figure 1b shows the typical final morphology of the selfassembled LNTs from 1 taken with the conventional TEM (JEM2000EX). The average inner and outer diameters of the LNTs were 61 nm (standard deviation (S.D.): 9.9 nm) and 400 nm (S.D.: 90 nm) at that cooling rate, respectively. Figure 2 shows the initial structure of the lipid nanotubes observed using ZPC-TEM. An edge structure associated with a growing lamellar sheet was observed only at the earliest stage of the self-assembly. The ZPC-TEM image allowed us to evaluate the width of the sheet as ca. 250 nm. This figure suggests the extending lamellar sheet rolls up to the tube morphology along to the tube axis. A tube structure grown from an edge structure is shown in Figure 3a. Although the lengths of the observed core tubes showed some distribution, the inner and outer diameters were quite uniform for each tube, giving 130 and 180 nm, respectively. It is a sharp contrast to those of final nanotube structure with a broad distribution of the diameters as mentioned above. We denote the initial, uniform nanotube structure as a “core tube”. At the tip of the core tube, the growing sheet-like structure can also be observed. Figure 3b shows the magnified image of the dashed square in Figure 3a. We can observe a helical mark indicating the boundary of the neighboring lamellar tape, as well as clear layer structures in the wall of the core tube. One layer will correspond to a single bilayer of the lipid molecules.12,19 In this ZPC-TEM image, five sheets of bilayers were clearly observed to constitute the wall of the core tube with rolling-up. These ZPC-TEM images allowed us to determine the sheet width as 250 nm and pitch angle of the winding as 45°. This value gives important information to discuss the molecular packing in the lamellar sheet. In addition, for all the observed core tubes, no gap was observed between the lateral sides of the neighboring lamellar tapes (the part indicated by the white arrows).
Growth Process and Molecular Packing of LNTs
Figure 3. (a) ZPC-TEM image corresponding to the growth of the core lipid nanotube. Growing edge with lamellar sheet can be seen. (b) Enlarged image of the core lipid nanotube. The white arrow indicates a helical mark showing that the core tube is formed by rolling-up of the lamellar sheet side by side. The wall of the core tube consists of five bilayer sheets. Scale bars in (a) and (b) indicate 100 nm in length.
Two typical routes are well-known for the nanotube formation from a helically coiled ribbon to a tubular structure.1 One route proceeds with shortening of the helical pitch of a ribbon structure with a constant tape width.15,16 The other proceeds with widening of the tape width with a constant helical pitch.14 Both processes involve intermediate structures with gaps between the neighboring lamellar tapes. For all the present core tubes, however, the sheet width (250 nm) and the pitch angle (45°) were constant and no gap was observed. This result might demonstrate that the core tube was directly formed via neither route mentioned above. The direct formation of the gapless nanotube geometry is also evidenced by the following estimation: the sheet width (250 nm) and the pitch angle of 45° fits well the tube outer diameter of 180 nm ()the sheet width multiplied by the cosine of the pitch angle). This explains why the core tube structure was very rapidly formed at temperatures just below the gel-to-liquid-crystalline phase transition. After the rapid formation of the core tube with uniform inner and outer diameters, it grows into a thicker nanotube structure (15) Chung, D. S.; Benedek, G. B.; Konikoff, F. M.; Donovan, J. M. Proc. Natl. Acad. Sci. U.S.A. 1993, 90, 11341. (16) Shimizu, T.; Hato, M. Biochim. Biophys. Acta 1993, 50, 1147.
Langmuir, Vol. 24, No. 3, 2008 711
Figure 4. (a) ZPC-TEM image showing the growth of both inner and outer bilayer sheets around the core lipid nanotube. Scale bar in the figure indicates 50 nm in length. The bilayer sheet winds around the core tube at the angle of 45° with respect to the tube axis (the white arrows). (b) ZPC-TEM image of the layered structure of the lipid nanotube. Scale bar in the figure indicates 50 nm in length. Stacked seven lamellar sheets with 40 nm in thickness can be seen.
as shown in Figure 4a. At this stage, the lamellar sheets stacked on both the inner and outer surfaces of the core tube. Interestingly, clear boundaries of the additional stacking sheets are observable (the white arrows in the figure) on the outer growing surface. Figure 4b shows a magnified image of the wall of a typical growing LNT. We were able to count seven layers and the thickness of the wall was 40 nm. Thus, the ZPC-TEM image enabled us to evaluate the thickness of one bilayer sheet as ca. 4.6 nm, corresponding to those observed in Figure 3b. About the curved membrane structure, Schnur and co-workers have characterized phospholipid nanotubes by circular dichroism spectroscopy and evidenced the curved membrane forms owing to the chiral lipid molecular packing.17 In our previous paper, we also measured the circular dichroism of the lipid nanotubes and found the chirality in the lipid packing is critical for the curved membrane morphology.12 Furthermore, Schnur and co-workers reported the relationship between the amphiphilic molecular structure and the molecular packing and nanotube dimensions in their phospholipid systems.18 For the present glycolipid, how does the chirality in the glucose head group transfer the chiral molecular packing in the membrane? We speculate that the chiral lipid packing is stabilized by the (17) Schnur, J. M. et al. Science 1994, 264, 945. (18) Singh, A.; Wong, E. M.; Schnur, J. M. Langmuir 2003, 19, 1888.
712 Langmuir, Vol. 24, No. 3, 2008
Yui et al.
Figure 5. X-ray diffractogram of the wet lipid nanotube. 2θ is the diffraction angle. λ ) 0.1542 nm (Cu KR line). Miller indices are indicated at the peaks.
Figure 6. IR spectrum of the CH2 scissoring bands of the lipid nanotube.
hydrogen bonding between lipid amide groups. In addition, we reported the saturated homologue of 1 does not form any nanotube morphology.12 This result strongly suggests the bend alkyl-chain conformation due to the cis CdC double bond is critical to the chiral packing in the hydrophobic part. To obtain deeper insight into the molecular packing of the lamellar sheets, we also measured X-ray diffraction (XRD) and IR spectra of the LNTs. Figure 5 shows the XRD pattern of the wet LNT. The small-angle region shows the strong (001) reflection peak and weak (004) and (006) ones. The ratio of the corresponding d-spacing values was practically 1:1/4:1/6, indicating a lamellar structure consisting of the stacked lipid bilayers. The lamellar repeat was evaluated as 4.54 nm, which are well compatible with that obtained from the ZPC-TEM image (4.6 nm). On the basis of the small-angle X-ray diffraction, we have derived the one-dimensional electron density profile of the lipid bilayer in the previous paper.19 In the wide-angle region, a sharp and a strong but broad peak were observed at 2θ of 20.2° and 21.5°, respectively (Figure 5). The d-spacing values of these two peaks are evaluated as 0.439 and 0.414 nm, respectively. It is noted that the observed pattern is of an orthorhombic perpendicular subcell (O⊥), a typical type of lateral chain packing in solid lipid membranes.20 It has been known that the δ (CH2) scissoring band sensitively reflects the molecular packing of lipid alkyl chains.21 Figure 6 shows the IR spectrum of the same sample in the δ (CH2) scissoring modes. The characteristic split peaks to the O⊥ subcell structure were observed at 1458 and 1468 cm-1. The IR spectrum also evidences the assignment by the XRD measurements. The assignment allowed us to derive the lattice constants (Figure 7). In the O⊥ subcell structure, the diffraction lines of 0.439 and 0.414 nm were assigned to (20)S and (11)S, respectively. We can evaluate the lattice constants of the orthorhombic subcell (19) Guo, Y.; Yui, H.; Minamikawa, H.; Masuda, M.; Kamiya, S.; Sawada, T.; Ito, K.; Shimizu, T. Langmuir 2005, 21, 721. (20) Ruocco, M. I.; Shipley, G. G. Biochim. Biophys. Acta 1982, 691, 309. (21) Lewis, R. N. A. H.; McElhaney R. N. In Infrared Spectroscopy of Biomolecules; Mantsch, H. H., Chapman, D., Eds.; Wiley-Liss, Inc.: New York, 1996; Chapter 7.
Figure 7. (a) Schematic illustration of the molecular packing in the subcell structure of the LNT. (b) and (c) Schematic illustrations of the molecular packing of the subcell structure in a and b axes presented in (a), respectively.
structure, as ) 0.878 nm and bs )0.466 nm from the following equation,
1 h2 k2 ) + Dhk2 as2 bs2 where h and k are Miller indices in the subcell structure and Dhk is the d-spacing value. Because the LNT is made of an interdigitated bilayer structure,12,19 the orthorhombic unit subcell structure contains two alkyl chains. Thus, we can evaluate the occupied area of one alkyl chain as 0.20 nm2. This area value calculated from the present assignment is appropriate for the area of the monoene alkyl chain in a solid state.22 The observed chain interdigitation was rationalized by a cross-sectional area of glucose head group (typically 0.45-0.50 nm2) and that of alkyl chain (0.20 nm2). Hydrogen bonds between the neighboring lipid molecules play an important role in the formation of lamellar sheet and tubular structure. The length of the hydrogen bond between the amide (NH) and carbonyl (CO) groups is within 0.48-0.49 nm.23 In the evaluated lattice model, the nearest neighboring lipid molecules in the bs axis meet this requirement. Thus, in the bs (22) Abrahamsson, S.; Nahringbauer, I. R. Acta Crystallogr. 1962, 15, 1261. (23) Leiserowitz, L.; Tuval, M. Acta Crystallogr. 1978, B34, 1230.
Growth Process and Molecular Packing of LNTs
axis, the successive hydrogen bonding between the NH and CO groups develops, resulting in the extension of the lamellar tape. On the assumption that the neighboring distance between the lipids molecules in the horizontal direction of the lamellar tape is 0.485 nm (the average of 0.48-0.49 nm), we can evaluate the tilt angle of the lipid molecules is ca. 16° to the bilayer normal. Chiral tilted bilayer model has been widely employed to describe the chiral packing self-assembly of lipid nanotubes.14,24-26 In this model, two parameters are critical factors in controlling the size and shape of LNTs: a molecular tilt and a twisting angle. The molecular tilt in the lipid membrane is closely related to the growth anisotropy of the LNT, and the packing at nonzero angle between the neighboring molecules generates spontaneous curvature of the lamellar sheet, leading to the nanotube diameter. However, it has been difficult to determine these factors experimentally. Recently, Fang and co-workers successfully visualized the molecular tilt direction of chiral phospholipid tubules (0.7 µm width) by means of liquid-crystal imaging.27 Here, we estimate the twisting angle between the neighboring molecules in the LNT walls by combining the information from the ZPC-TEM image and XRD results. The ZPC-TEM images of the core tube (Figure 3 and 4) allowed us to estimate its inner and outer diameter to be 130 and 180 nm, respectively, and the pitch angle of the lamellar tape, 45°. When considering that the diameter is 150 nm, we can induce the length of the lamellar tape for one roll around the tube axis as 666 nm ()150 nm × π (3.14)/cos 45°). The number of molecules in one periodic lamellar tape can be estimated as 666/0.485 ) ca. 1400. Thus, the twisting angle of the neighboring lipid molecules can finally be evaluated as about 0.26° (360°/1400) (Figure 8). Here, we assumed uniform tilt for the present small nanotube for simplicity. This value is similar to that of the twisting angle of cholesteric liquid crystals.28 Evaluating twisting angles in lipid nanotubes would be useful in understanding the chiral packing of the lipid molecules and designing nanotube morphology. (24) Ou-Yang, Z.; Liu, J. Phys. ReV. Lett. 1990, 65, 1679. (25) Selinger, J. V.; MacKintosh, F. C.; Schnur, J. M. Phys. ReV. E 1996, 53, 3804. (26) Koumura, S.; Ou-Yang, Z. Phys. ReV. Lett. 1998, 81, 473. (27) Zhao, Y.; Mahajan, N.; Lu, R.; Fang, J. Proc. Natl. Acad. Sci. U.S.A. 2005, 102, 7438. (28) de Gennes, P. G.; Prost, J. The Physics of Liquid Crystals, 2nd ed.; Oxford University Press: Oxford, 1983.
Langmuir, Vol. 24, No. 3, 2008 713
Figure 8. Schematic image of the twisting angle (φ) in chiral packing by the neighboring lipid molecules. φ is evaluated as 0.26° from ZPC-TEM image, XRD pattern, and IR spectrum measurements.
Conclusion Combination of the ZPC-TEM microscopy and quick freezing sampling technique enabled us to measure the initial structures of a self-assembled lipid nanotube. We found that the growing process of the lipid 1 consists of rapid formation of the core tube by rolling-up of a lamellar tape without gaps and the subsequent one-by-one piling-up of lamellar sheets around the inner and outer surfaces of the core tube. In addition, the helical pitch observed using the ZPC-TEM images, XRD diffraction pattern in the wide-angle region, and IR spectrum enabled us to discuss molecular packing of the LNT. The twisting angle of the LNT was, thus, determined for the first time. These results should provide new information on the relation between the molecular structures and morphologies of self-assembled lipid nanotubes. Acknowledgment. Dr. Mitsutoshi Masuda (NARC, AIST) is acknowledged for the discussion on the assignment of the IR spectrum and providing numerical data of the single crystal of N-alkanoyl-β-D-glycosylamines. Mr. Hideo Tsunakawa, Mr. Toshio Ito, and Mr. Yasutomi Kakegawa (Institute of Engineering Innovation, School of Engineering, University of Tokyo) helped us operate the conventional TEM microscope. LA702488U
