Biochbnica et Biophysica Acre, ! 127 ( i 992) 157-162 © 1992 Elsevier Science Publishers B.V. All rights reserved 0005-2760/q2/$05.0~}

157

BBALIP 53983

Observation of the inverted hexagonal phase of lipids by scanning tunneling microscopy K. Zeng and K.C. Lin Department of Biophysics, Beijing Medical Uni'ersity, Beijmg (China) (Received 28 April 1992)

Key words: Scanning tunneling microscopy; Inverted hexagonal phase of lipids; Lipid phase

Scanning tunneling microscopy (STM) has been used in direct observations of the inverted hexat:onal phase, of several iipids, such as cardiolipin (CL) in the presence of Ca 2+, soybean PE, DOPE and a binary system of CL and DMPC as well. Single tube-like structures, bundles of tubes and the cross-sectional view of the H . phase can clearly be seen at room temperature in a near natural state of these lipids. The success of this study shed some light on the applicability of STM in the investigation of lipid polymorphism and its role in the function of biological membranes.

Introduction

The discovery of the polymorphism of lipids has attracted much attention in the last 15 years since the transition between bilayer-to-non-bilayer structures is a problem of importance in the elucidation of the function of biomembranes. The most common type of nonbilayer conformation of lipids is the inverted hexagonal (H u) phase (a tubular structure), which has been identified mostly by X-ray diffraction [1], 31p-NMR [2] and freeze-fracture electron microscopy [3]. Indeed, the progress in this field has been much promoted by the successful applications of these techniques. However, each of them has its own drawbacks either because the sample state was far from natural, or because only indirect information could be drawn with some ambiguities. Much efforts have been made in recent years, using differential scanning calorimetry [4], infrared and Raman sp6ctroscopy [5,6], absorption spectroscopy [7], fluorescence polarization and fluorescence spectroscopy [8,9], and also some time-resolved techniques [10,11], in the hope to identify at an improved resolution or to study the dynamical process of the transition.

Correspondence to: K.C. Lin, Department of 13i~physics, Beijing Medical University, Beijing, China 100083. Abbreviations; STM, scanning tunneling microscopy; TEM, transmission electron microscopy: CL, cardiolipin; PE, phosphatidylethanolamine; DOPE, dioleoylphosphatidylethanolamine; DMPC, dimyristoylphosphatitylcholine; IMP, intramembranous particles; PBS, phosphate-buffered saline.

The invention of scanning tunneling microscopy (STM) by Binnig and Rohrer [12], which was laureated with the Nobel Prize in 1986, provided a new powerful tool tor the surface investigations of metals and semiconductors, and biological macromolecules and binmembranes as well [13,14]. On the basis of our own work in the study of Langmuir-Blodgett membranes, reconstituted proteoliposomes and living cells [15,16], we tried to use STM in observing the H~l phase of various lipids and mixtures thereof. Materials and Methods

DMPC, CL, DOPE and soybean PE were purchased from Sigma. All other reagents were of analytical grade. Usually, 5 mg lipid in 2 mi chloroform/methanol (2:1 v/v, and in all cases thereafter)was dried by evaporation in a small vessel for 3-4 h and kept overnight under vacuum. 1 ml phosphate-buffered saline (PBS) (pH 8.0) was added and the suspension was sonicated with a bath sonicator for 5'-20'. For pure CL, 0.6 ml CHCla/CH3OH plus 0.2 ml double distilled water was added. After sonication, the sample was divided into two aliquots, to one of them CaCl 2 was added where C L / C a 2+= 1 (M/M). In the case of lipid mixture, 12 mg CL and 3 mg DMPC were dissolved in 2 ml organic solvent, and after evaporation 2 mi Tris-HC1 with 1% EDT'A (pH 7.2) was added. By using a micropipet, the sample was transferred onto the substrate, which was a piece of cover glass gilded by evaporation of gold and then quenched, in

158

Observations were made with the following parameters: Vbi~ = 100-150 mV with tip negative, a n d lli p = 1.0-1.2 nA. Either amplitude modulation line scan image, brightness modulation top view image, or shaded topographical image (the mixture of the above two images) could be obtained on the oscilloscope. Photographs of these images were further processed by an SEM-IPS Image Analyzer (Cantron, Germany). Four kinds of processed images can be obtained: high resolution grey-scale image, three dimensional stereo processed image, contour plot image and line scan image transformed from brightness to height.

order to observe the topographical feature of the sample in solution, water with 50% glycerol was placed on the sample. To improve the resolution of the image, a gold layer of 1.0-1.5 nm thick can be sprayed at normal incidence (as opposed to oblique incidence used in TEM) on the dried surface of the sample. STM images were obtained by the use of an SSX-I Scanning Tunneling Microscope. designed by the Central Laboratory of Electron Microscopy, Academia Sinica [17]. Lateral resolution in x- and y-directions was 0.1-0.2 nm, and that of z-direction was 0.01 nm. The tip used was electrochemically etched tungsten wire 0.5 mm in diameter. Tips used in imaging the objects were carefully selected by trial-and.error. Images were displayed on an oscilloscope, z-direction calibration was done by a three-position momentary toggle switch. The switch can add an equivalent z drive voltage of 5 or 0.5 V to the scope display. From the deflection of the trace on the scope and the sensi. tivity of the microscope head one can calibrate the height scale of the image in nm/V. in our case, 5 V and 0.5 V correspond to 10 nm and 1.0 nm, respectively.

Results

Obsen'ation of tlze H n robes of CL / Ca" ÷ To observe the fine structure of H , tubes without any additional treatment of the sample, a system of C L / C H C i 3 / C H 3 O H / C a 2+ was selected. When a sonicated sample was spread on the substrate and the organic solvents evaporated, H u tubes will appear as observed using STM (see Fig. la). However, when a considerable proportion of water was present, only the

F•8.

|, ....

•"

" ..................

b

---,,11u,-,s~.,1,.

,~ s " r j

• •--

y,

u. | I

lrltL.

ILdl

IL.,IL. " t " C a

2 +

covered with .50% glycerol. (b) CL+Ca z+ directly after air-drying. (c) CL+Ca -~+ alter air-drying and covered with gold. (d) CL without Caz+ after air-drying.

159

Fig. 2. STM observation of H n structures of soybean PE. (a) Soybean PE with 50~, glycerol after air-drying, x = y = 100 nm. (b) Cross-sectional view of the tubes, x = y = 15{)nm.

outer surface of bundles of tubes, i.e., envelope of the tubes could be seen. Along with gradual evaporation of free water, the fine structure will become clear so that individual tubes or even single tubes will appear (Fig.

lb). The outer diameter of these tubes was 6.0-8.0 nm, in good accordance with those fi'om X-ray diffraction studies [18]. For a further improvement of the sharpness of the image, a thin layer of gold was sprayed on

/

~

~ !i!ii~ : i~ i!~i~i!~! i~iii!iii !i ~

i

. . . .

Fig. 3. High resolution images of soybean PE. (a) Original grey scale image of PE coverd with gold. (b) Stereo porcessed image of a, in both pictures x -- y -- 160 nm. (c) Original grey scale image of lhe cross-sectional view, x -- y -- 35 nm. (d) Stereo processed image 'ff c, x = y -- 3.'; nm.

ing the process of drying. In fact shrinkage was also evident in the former type of image as can be seen from Fig. 3b, where the heights of these tubes were estimated to be 2.0-4.0 nm. It should be noted that both drying and flattening occurred in the tubes lying horizontally. Besides, one can see that small fluctuations in tube diameters are evident, these may be due to the formation of intramembranous particles (IMP). Simillar results were obtained with DOPE.

the dried sample. The result was shown in Fig. lc. On the other hand, a sample without any Ca 2+ adopted the large lameUar structure as shown in Fig. ld.

Obsen,ation of the H , tubes of soybean PE and DOPE Both fully hydrated soybean PE and DOPE will adopt H u structure at room temperature since the transition temperatures of these two lipids from bilayer to H u phase are -10 and + 10°C, respectively. The H u structure of lipids can give two different types of images as shown in Fig. 2a and b. Fig. 2a shows a shaded topogrdphical image of soybean PE under 50% glycerol and water. This is the tubular structure usually seen with the tubes lying horizontally on the substrate. This can be further proved with the processed images as shown in Fig. 3a and b. The other type of image from the H u structure is like in Fig, 2b. In this case one can only see the white spots arranged partially in a hexago. nal form, A small area in this picture was further processed to give clearer pictures as shown in Fig. 3c and d, it is clear from these images that they represent the cross-section of the H~, tubes, with their internal and external diameters as 5.0 and 7.0 nm, respectively. These are slightly smaller than those obtained with fully hydrated sample, because shrinkage occurred dur-

Obserz'atiot~ of DMPC / CL mb:tures In a mixed system of CL/DMPC (2: l, M / M ) without Ca -'+, a bilayer structure will result. Fig 4a shows the STM of a boundary between two liposomes. However, in the presence of Ca 2+, phase separation will occur as shown in Fig. 4b. In order to improve the resolution of these images, samples were covered with gold after drying. Fig. 4c shows the surface structure, in which two types of structures can be seen: one is tubular, in a parallel arrangement. The diameters of these tubes are approx. 8.0-10.0 nm. The other is a rather flat surface of the lamellar region. More interestingly, we can clearly see the interface between these two phases as indicated by the arrows in the figure.

mPcp

~i

I

Fig. 4. STM images o f C L / D M P C ( 2 : 1 , M / M ) . ( a ) S h a d e d topographical image without Ca 2+. x = y = 100 nm. (b) Grey-scale image with Ca 2+, x -- y -- 100 nm. (c) High resolution image covered with gold, x = y = 80 nm.

161

Fig. 5. Imagesof sonicated liposomesof CL/DMPC. (a) Grey-scale image without Ca2+, x = y = 120 nm. (b), Same sample after adding Ca2., x = y = 120 nm. (c) Freeze-fractureelectron micrographof the sample in a. (d) Freeze-fractureelectron micrographof the sample in b.

Sonicated liposomes have also been imaged after drying and sprayed with an Au layer (see Fig. 5a and b). In the absence of Ca 2+, liposomes are in bilayer f~rm with smooth surfaces in a bell-like shape. However, because of drying and covering with gold, they were flattened. Their diameters are in the range of 60.0-120.0 nm. While in the presence of Ca 2+, smooth surfaces of the bilayers were dertroyed, and there appeared many small particles with diameters of 8.010.0 nm and height of approx. 2.0 nm (Fig. 5b). These are thought to be intramembranous particles as seen usually in freeze fracture replicas by TEM shown in Fig. 5c and d.

Discussion Although the exact theory for STM of biological objects is not well established yet, the fact that the H n structure of some non-bilayer forming iipids will result under some conditions gave us a clue for testing the applicability of the STM technique in direct observation of the H n structure. By careful examination of the substrate itself and the control system, we succeeded in obtaining the evidence for the formation of H u struc-

ture by these lipids and a mixture containing CL and DMPC. From this work, we got some experience in using STM to image structures of lipid assemblies. Air-drying will remove most of the free water, but not the bound water of the membrane surfaces, so it is the most frequently used method. However, 50% glycerol/water could be added on the surface of the sample, so that the state will in a more natural one which is desirable in imaging living cells [16]. In this case, the resolution will be lower compared to that of the dried one. However, by careful comparison of these two kinds of images, or after covering a thin layer of gold, reliable information could be obtained. The dryine process usually reduces the dimension of some details of ttle sample, e.g., the H it-tube diameter, and especially the flattening of these tubes when they are lying horizontally on the substrate as in Fig. 3b. In this case, the image of the cross-section of the tubes will give us more reliable information. The reason why some bundles of tubes are lying on the substrate (longitudinally) whilst others are standing on the substrate (tranversally, so that we can see the cross-section of a bundle of tubes) is primarily depending on the lengths

15 of these tubes. In the process of settling down during drying, long tubes will adopt the former orientation while the shorter ones will stand on the substrate. In fact these phenomena have already been investigated by Caron [18] and Stoekenius [19] with TEM and X-diffraction, respectively, and their result supported our explanation of these images. Fluctuations in the diameters of Hit tubes are frequently seen in soybean PE and CL/DMPC binary system. ~ i s may be resulted from the formation of IMP in liposomes. We have used sonicated liposomes and got images by both STM and TEM, as shown in Fig. 5a-d. We are convinced that IMP is a common type of non-bilayer in liposomes and tubes are formed by close arrangement of these IMPs. It is our belief that STM may play an important role in the investigations of lipid polymorphism. By extrapolating the research to more complex systems, STM may be useful in studying the surface of cell membranes as well. Some work of this kind is currently undergoing in our laboratory.

Acknowledgements We are grateful to Prof. J.E. Yao and his group in the Central Laboratory of Electron Microscopy, Academia Sinica for technical assistance, and also to Prof. Y.X. Su, BMU for invaluable comments and discussions. This work was supported by the National Natural Science Foundation of China.

References 1 Luzzati, V. (1968) in Biological Membranes, Vol, 1 (Chapman, D., ed.), pp. 71-123, Academic Press, London. 2 Cullis, P.R. and De Kruijff, B. (1979), Biochim. Biophys. Acta 551,399-420. 3 Gulik-Krzywicki, T. (1975) Biochim, Biophys. Acta 415, !-28. 4 Van Dijk, P.W.M., De Kruijff, B,, Van Deenen, LL.M,, De Gier, J, and Demel, R,A, (1976) Biochim. Biophys, Acta 455, 576-587. 5 Mantsch, H,H., Martin, A. and Cameron, D.G. (i981) Biochemistry 20, 3138-3145. 6 Levin, I.W. (1984) in Advances in Infrared and Raman Spectroscopy, Vol, 11 (Clark, R,J.H. and Hester, R.E., eds.) pp. 1-48, Wiley Heyden, New York, 7 Veiro, J.A., Khalizah, R.G, and Rowe, E.S, (1989) Bioch,n. Biophys. Acta 979, 251-256. 8 Cheng, K.H. (1989) Biophys. J. 55, 1025-1031, 9 Hong, K., Baldwin, R.L. and Papahadjopoulos, D. (1988) Biochemistry 27, 3947-3955. 10 Caffrey, M (1987) Biochemistry 26, 6349-6363. II Siegel, D.P., Burns J.L., Chestnut, M.H. and Talmon, Y. (1989) Biophys. J. 56, 161-169. 12 Binnig, G., Rohrer, H., Gerber, C.H. and Webel, E, (1982) Appl. Phys. Lett. 40, 178-180. 13 Zasadzinski, J.A.N., Schneir, J., Gurley, J., Elings, V. and Hansma, P.K. (1988) Science 239, 1013-1014. 14 Binnig, G. (1986) Bull. Am. Phys, Soc. 31,217. 15 Zeng, K., Lin, K,C. and Yao, J.E. (1991) Biophys. Acta Sinica 7, 476-480. 16 Dai, J.W., Jiao, Y.K,, Dong, Q., Su, Y.X. and Lin, K.C. (1991) J. Vac. Sci. Technol. B 9, 1184-1188. 17 Yao, J.E., Shang, G.Y., Jiao, Y.K., Bai, C.L., He, J., Zhong, J.C. and Rong, D.N. (1988) J. Microsc. 152, 671. 18 Caron, F., Mateu, L., Rgny, P. and Azerod, R. (1974) J, Mol. Biol. 85, 279-300. 19 Stoekenius, W. (1962) J. Cell Biol. 12, 221-229.