Two-chamber AFM: probing membrane proteins separating two aqueous compartments

Rui Pedro Gonçalves¹, Guillaume Agnus², Pierre Sens³, Christine Houssin⁴, Bernard Bartenlian² & Simon Scheuring¹

¹ Institut Curie, UMR168-CNRS, 26 Rue d'Ulm, 75248 Paris, France.

² IEF, MMS, Université Paris-Sud, Bat. 220, 91405 Orsay, France.

³ ESPCI, CNRS-UMR 7083, 10 rue Vauquelin, 75231 Paris, France.

⁴ LBMII, IGM, Université Paris-Sud, Bat. 360, 91405 Orsay, France.

Biological membranes are 40 Å–thick layers consisting of lipids and membrane proteins. Cell membranes define the ultimate inside and outside of all living cells. Eukaryotic cells feature intracellular compartments equally confined by membranes. Functional studies on membranes and membrane proteins within them are performed by several techniques that preserve the native concept of membranes that separate two aqueous compartments: patch clamp¹, black lipid membrane² and spectroscopy on membrane vesicles³. Structural studies on membrane proteins, in contrast, observe rather artificial systems. Membrane-protein high-resolution structures are solved by X-ray diffraction of detergent-solubilized, purified and three-dimensionally crystallized proteo-detergent micelles⁴. Cryo–electron microscopy of two-dimensionally crystallized membrane proteins has rarely resulted in atomic-resolution structures⁵. Single-particle electron microscopy analysis has yielded structures at medium resolution⁶. In both cases, membrane proteins are solubilized, purified, crystallized, cryo-fixed in ultra-high vacuum, support-adsorbed or subjected to several of these non-native preparation conditions. AFM⁷ has developed into a powerful tool in membrane research. Topographs at 10-Å resolution⁸ have been acquired on two-dimensionally crystallized membrane proteins⁹ and, most recently, on native membranes¹⁰ containing the supramolecular architecture of multiple membrane protein complexes¹¹. Although these were close-to-native conditions, the membranes were nevertheless support-adsorbed.

We performed high-resolution imaging on three types of holey Si(001) surfaces, with different hole diameters and periodicities (90 nm and 200 nm, 150 nm and 500 nm, 250 nm and 500 nm, respectively). Supports with small holes and short hole periodicities facilitate membrane imaging. Before membrane adsorption, we performed extensive AFM-support characterization (Supplementary Fig. 2 online). To develop our approach, we chose the native S-layer membrane of C. glutamicum, because (i) it can be produced in large quantities¹⁴, (ii) we have investigated its topography¹⁵ and (iii) intermolecular forces¹⁵, (iv) its periodic structure facilitates image processing¹⁵, (v) because it is a native membrane, and finally (vi) because of its known stability¹⁶. We successfully adsorbed membranes to the holey Si support using buffers containing divalent ions. We often found flatly adsorbed membranes with sizes larger than 500 nm that completely cover holes (Fig. 1a). We repeatedly imaged individual membranes containing nonsupported membrane areas (Fig. 1b). We detected, at intermediate magnification, the two-dimensional (2D) lattice of the S-layer proteins in deflection images (Fig. 1c). Section analysis across the membrane with underlying holes (Fig. 1d) showed that nonsupported membrane areas withstood the force applied by the AFM tip. Noncovered holes appeared substantially deeper than the membrane thickness, even though we adjusted scan parameters for membrane imaging, that is, minimal loading forces (100 pN) and relatively fast scan speed (20 m/s). The membrane thickness was 4.5 nm, in agreement with published results¹⁵. For better comparison, we analyzed profiles in scan direction, proving preservation of nonsupported membrane areas (Fig. 1e).

Figure 1. Membrane adsorption on holey Si(001) surfaces. (a) Overview topograph of S layers adsorbed to the holey Si(001) AFM support. Many membranes cover several holes (arrows). (b) Medium-magnification topograph of the S layer outlined in a. (c) Deflection image of b revealed the hexagonal crystalline packing of the S-layer proteins. (d) Section analysis along the red line (m) in b showed the membrane and the hole profile. (e) Section analysis along the white lines (1–9) in b revealed profiles of noncovered and covered holes. All section profiles are represented leveled to their immediate environment; in reality the surfaces of section analysis 5 and 6 are on top of the membrane (4.5 nm above the silicon surface and above the section graphs 1–4 and 7–9). Scale bars, 2 m (a) and 500 nm (b,c). Full Figure and legend (36K)

We imaged nonsupported membranes using cantilevers with spring constants of 0.20 N/m, applying the rationales of electrostatic balancing of tip-sample interactions⁸. According to these, a sharp local tip of 2.5 nm radius¹⁷, which allows high-resolution membrane-surface contouring at van der Waals distance, protrudes from a global tip that distributes the force applied by long-range electrostatic interactions over large sample areas⁸. Probably such long-range electrostatic balancing rendered imaging of nonsupported membranes possible by redistributing parts of the loading force to the surrounding supported membrane areas. We imaged the inner (Fig. 2a) surface of the C. glutamicum S layer on nonsupported membrane areas at high resolution. The unit cell dimensions were a = b = 16 nm, = 60 °. Notably, we visualized the subunit architecture of the S-layer proteins in the raw-data topographs (Fig. 2a) and Fourier-filtered images (Fig. 2b). We could enhance features through averaging and subsequent symmetrization (Fig. 2c). We equally imaged and analyzed the outer surface of the C. glutamicum S layer (Fig. 2d–f). Topographs of the nonsupported membrane areas compared very favorably to the data obtained by high-resolution AFM analysis of supported membrane layers¹⁵. Notably, we unambiguously visualized in raw-data topographs and, in averages, structural details such as the 15-Å diameter pore in the center of the triangle structure on the outer surface (Fig. 2f), comparable in size to channel pores¹⁸.

Figure 2. High-resolution imaging of nonsupported membranes. (a) High-resolution topograph of the inner S-layer surface (outline: hole position with 90-nm diameter). (b) Fourier-filtered image of a. (c) Average (left) and symmetrized average (right) topographs. (d) High-resolution topograph of the outer S-layer surface (outline: hole position with 200-nm diameter). (e) Fourier filtered image of d. (f) Average (left) and symmetrized average (right) topographs. Outlines in c and f delineate the structural pore in the center of the flower (F) and triangular (T) shaped surfaces of 15 Å in diameter. Dashed outlines in c and f delineate the structural pore in the center of the flower (F) and triangular (T) shaped surfaces of 15 Å in diameter. Scale bars, 90 nm (a), 20 nm (b), 10 nm (c), 50 nm (d), 20 nm (e) and 10 nm (f). Full Figure and legend (54K)

We could repeatedly image membranes containing nonsupported areas using minimal forces (200 pN) applied to the AFM tip. The periodic arrangement of the holes in the silicon surface allowed for precise determination of the position of membrane-covered holes (Fig. 3a). To probe physical parameters of nonsupported membranes, we performed force measurements on these areas. Imaging (Fig. 3b) and profile analysis (Fig. 3c) after the force measurement corroborated successful piercing of the nonsupported membranes. Force-distance curves on nonsupported membrane areas had an unusual, but expected, shape (Fig. 3d). As indicated by the approach curves, after initial contact of the tip with the membrane, the membrane was indented until rupture. This process allowed the cantilever to relax, that is, unbend, until maximum depth. It was notably this snap-to-contact after rupture that unambiguously reported the piercing event of nonsupported membranes. Furthermore, the total distance between the initial contact point with the membrane, and the final surface contact of 50 nm matched well with the depth measured in holes of this support (Supplementary Fig. 2). The retract curves corroborated that the tip at zero-force level could be below the initial contact point. On further retraction, adhesion forces occurred probably owing to membrane parts that stuck to the tip.

Figure 3. Membrane piercing. (a,b) Medium-magnification topographs recorded before (a) and after (b) the membrane-piercing experiment. Scale bars, 500 nm. (c) Section analysis along the white lines in b. The nonsupported membrane area covering the central hole initially undistinguishable in flatness from the rest of the membrane was pierced during the experiment. Both section analyses (profiles 1 and 2) documented an indentation far deeper than the silicone surface (yellow bars). (d) Two typical force-distance cycles recorded upon piercing nonsupported membranes. Initial tip-membrane contact (arrow 1), followed by indentation and membrane piercing (arrow 2). Then the cantilever snapped back to a more relaxed state where it came into solid contact (arrow 3). Upon retraction, a zero-force cantilever position (arrow 4) was detected below initial membrane contact (compare positions indicated by arrows 4 and 1). Bottom, control curve obtained on supported membrane. (e) Membrane piercing force-separation curve overlays. Curves were aligned on initial membrane contact, set to the graph origin. The point cloud of superposed curves was well fitted by a cubic power-law (red line). Inset, curves were aligned on the piercing-event peak. Membrane piercing forces determined between rupture event and force-curve baselines (dashed outlines; 1). Full Figure and legend (45K)

To relate the normal force F applied by the AFM tip to the lateral stress within the membrane, we assumed a linear response characterized by a 2D constant (energy per unit area) that we call 2D Young's modulus (E_2D). The stress is then related to the surface area expansion S of the total surface S by = E_2D (S/S)¹⁹. The area expansion increases with the square of the indentation depth h, within a numerical factor that depends on the geometry of the AFM tip. For a hemisphere it is S = h². The elastic energy stored inside the membrane is the integration of the stress over the surface expansion: = 0.5 SE_2D(S/S)² and the normal force F on the membrane is the derivative of this energy with respect to normal displacement h. It varies nonlinearly as the cube of the indentation: F = 2E_2Dh³ / L², where L = 75 nm is the hole radius. The data fitted well with such a cubic power-law (Fig. 3e) and gave access to the 2D Young's modulus E_2D = 0.2 N/m. This value corresponds well to the 3D Young's modulus measured for Murein sacculi²⁰ considering the thickness of the S layer of 4.5 nm and is of the same order of magnitude as the stretching modulus of a lipid bilayer²¹. We calculated that the bending rigidity¹⁹ of the S layer (0.1 E_2Dd²; 90 k_BT; where d is the layer thickness and T is the absolute temperature) is larger than the bending modulus of a lipid bilayer²¹, probably because fluid lipids reorganize upon stress and proteins cannot. Rupture occurs at 25 nm indentation (Fig. 3e), for which the area expansion is 10%, and a yield-force F = 3 nN. In contrast, as previously measured¹⁵, pulling on surface-adsorbed membranes revealed subunit rupture at 0.26 nN. In the present work, the membrane remained close to the linear elastic regime up to rupture (Fig. 3e). The elastic energy stored in one S-layer unit at rupture is 0.5 sE_2D(S/S)² 60 k_BT (where the unit cell area s = 221 nm²), corresponding to an energy of 10 k_BT per protein-protein contact. Our results suggest fairly weak protein-protein interactions in the S layer, the stability of which is assured by the cooperativity of units within the network.

To prove the functionality of hole-covering membrane proteins, we adsorbed purple membranes from H. salinarium, densely packed with the proton pump bacteriorhodospin, on the nano-indented substrates. AFM imaging (Fig. 4a) showed that membranes completely covered holes, and therefore created closed chambers with volumes of 10 attoliters. The pH-sensitive fluorescence probe 8-hydroxy-1,3,6-trisulphonic acid (pyranine)²² present in the adsorption buffer remained trapped in the chambers that are membrane covered, but we removed uncaptured pyranine by rinsing. Pyranine is highly sensitive to pH-changes around pH 7.2 (ref. 23), therefore reporting the occurrence of proton pumping by bacteriorhodospin into and out of the closed chambers, depending on the membrane orientation.

Figure 4. Proton pumping of nonsupported purple membranes. (a) Overview AFM topograph of purple membranes covering several chambers (top). Completely covered chambers (yellow arrows) enclose pyranine; upon rinsing, pyranine in noncovered holes (white arrows) was removed, and such holes did not give fluorescence response (bottom). Scale bars, 2 m (top) and 500 nm (bottom). (b) Sequence of green fluorescence images over 500 s documenting fluorescence increase (outline 1) and decrease (outline 2). Scale bars, 10 m. (c) Fluorescence intensity measurements reporting pH increase and decrease in covered chambers as function of excitation time of bacteriorhodospin. Full Figure and legend (26K)

	Top

	Top

We operated a commercial Nanoscope-E AFM⁷ (Veeco) equipped with a vertical engagement 160 m scanner (J-scanner) and oxide-sharpened Si₃N₄ cantilevers (length 100 m; Olympus), in contact mode at ambient temperature and pressure. We determined cantilever spring constants with a picoforce AFM (Veeco) using the thermal noise method resulting in k = 0.20 N/m. For imaging, we applied minimal loading forces of 200 pN, at scan frequencies of 4–7 Hz using optimized feedback parameters. We immerged the Si(001) holey supports in 20 l of adsorption buffer (10 mM Tris-HCl (pH 7.2), 150 mM KCl, 25 mM MgCl₂), and subsequently injected 3 l of membrane solution at 0.1 mg/ml protein concentration into the buffer drop. After 2 h we rinsed the sample with recording buffer (10 mM Tris-HCl (pH 7.2), 150 mM KCl). We acquired force-distance curves at rates of 80 nm/s, after progressively placing the tip to the center of a nonsupported membrane in imaging mode.

We adsorbed H. salinarium purple membranes for 10 min in adsorption buffer (1 mM Tris-HCl (pH 6.9), 200 mM KCl, 20 mM MgCl₂, 20 M pyranine), and then carefully rinsed with the same buffer without pyranine. We used a Leica DMRXA2 for fluorescence microscopy equipped with a 40 immersion objective with a numerical aperture of 1.4, and commercial wavelength filters. We excited pyranine fluorescence at 460 nm and detected it at 510 nm (ref. 29). We activated bacteriorhodospin by illumination at 570 nm, preventing fluorescence quenching of pyranine during bacteriorhodospin activation. We used pyranine excitation and detection times of up to 10 s considering the very small amounts of molecules per chamber. We compared all fluorescence intensity measurements to background.

	Top

1. Neher, E. & Sakmann, B. Single-channel currents recorded from membrane of denervated frog muscle fibres. Nature 260, 799–802 (1976). | Article | PubMed | ISI | ChemPort |
2. Mueller, P., Rudin, D.O., Tien, H.T. & Wescott, W.C. Reconstitution of cell membrane structure in vitro and its transformation into an excitable system. Nature 194, 979–980 (1962). | Article | PubMed | ISI | ChemPort |
3. Huang, C. Studies on phosphatidylcholine vesicles. Formation and physical characteristics. Biochemistry 8, 344–352 (1969). | Article | PubMed | ChemPort |
4. Caffrey, M. Membrane protein crystallization. J. Struct. Biol. 142, 108–132 (2003). | Article | PubMed | ChemPort |
5. Fujiyoshi, Y. The structural study of membrane proteins by electron crystallography. Adv. Biophys. 35, 25–80 (1998). | Article | PubMed | ISI | ChemPort |
6. Frank, J. Single-particle imaging of macromolecules by cryo-electron microscopy. Annu. Rev. Biophys. Biomol. Struct. 31, 303–319 (2002). | Article | PubMed | ISI | ChemPort |
7. Binnig, G., Quate, C.F. & Gerber, C. Atomic force microscope. Phys. Rev. Lett. 56, 930–933 (1986). | Article | PubMed | ISI |
8. Müller, D.J., Fotiadis, D., Scheuring, S., Müller, S.A. & Engel, A. Electrostatically balanced subnanometer imaging of biological specimens by atomic force microscopy. Biophys. J. 76, 1101–1111 (1999). | PubMed | ISI |
9. Schabert, F.A., Henn, C. & Engel, A. Native Escherichia coli OmpF porin surfaces probed by atomic force microscopy. Science 268, 92–94 (1995). | Article | PubMed | ISI | ChemPort |
10. Fotiadis, D. et al. Atomic-force microscopy: Rhodopsin dimers in native disc membranes. Nature 421, 127–128 (2003). | Article | PubMed | ISI | ChemPort |
11. Scheuring, S. & Sturgis, J.N. Chromatic adaptation of photosynthetic membranes. Science 309, 484–487 (2005). | Article | PubMed | ISI | ChemPort |
12. Oesterhelt, D. & Stoeckenius, W. Functions of a new photoreceptor membrane. Proc. Natl. Acad. Sci. USA 70, 2853–2857 (1973). | Article | PubMed | ChemPort |
13. Danelon, C., Perez, J.P., Santschi, C., Brugger, J. & Vogel, H. Cell membranes suspended across nanoaperture arrays. Langmuir 22, 22–25 (2006). | Article | PubMed | ChemPort |
14. Chami, M. et al. Organization of the outer layers of the cell envelope of Corynebacterium glutamicum: A combined freeze-etch electron microscopy and biochemical study. Biol. Cell 83, 219–229 (1995). | Article | PubMed | ChemPort |
15. Scheuring, S. et al. Charting and unzipping the surface-layer of Corynebacterium glutamicum with the atomic force microscope. Mol. Microbiol. 44, 675–684 (2002). | Article | PubMed | ISI | ChemPort |
16. Sleytr, U.B. Basic and applied S-layer research: an overview. FEMS Microbiol. Rev. 20, 5–12 (1997). | ChemPort |
17. Scheuring, S., Levy, D. & Rigaud, J-L. Watching the components of photosynthetic bacterial membranes and their "in situ" organization by atomic force microscopy. Biochim. Biophys. Acta 1712, 109–127 (2005). | PubMed | ChemPort |
18. Chang, G., Spencer, R.H., Lee, A.T., Barclay, M.T. & Rees, D.C. Structure of the MscL homolog from Mycobacterium tuberculosis: a gated mechanosensitive ion channel. Science 282, 2220–2226 (1998). | Article | PubMed | ISI | ChemPort |
19. Landau, L.D. & Lifschitz, E.M. Theory of elasticity (Pergamon Press, New York, 1986).
20. Yao, X., Jericho, M.H., Pink, D. & Beveridge, T.J. Thickness and elasticity of gram-negative murein sacculi measured by atomic force microscopy. J. Bacteriol. 181, 6865–6875 (1999). | PubMed | ISI | ChemPort |
21. Evans, E. & Rawics, W. Entropy-driven tension and bending elasticity in condensed-fluid membranes. Phys. Rev. Lett. 64, 2094–2097 (1990). | Article | PubMed | ChemPort |
22. Kano, K. & Fendler, J.H. Pyranine as a sensitive pH probe for liposome interiors and surfaces. pH gradients across phospholipid vesicles. Biochim. Biophys. Acta 509, 289–299 (1978). | PubMed | ChemPort |
23. Seigneuret, M. & Rigaud, J.L. Use of fluorescent pH probe pyranine to detect heterogeneous directions of proton movement in bacteriorhodopsin reconsituted large liposomes. FEBS Lett. 188, 101–106 (1985). | Article | ChemPort |
24. Stock, D., Leslie, A.G. & Walker, J.E. Molecular architecture of the rotary motor in ATP synthase. Science 286, 1700–1705 (1999). | Article | PubMed | ISI | ChemPort |
25. Oesterhelt, F. et al. Unfolding pathways of individual bacteriorhodopsins. Science 288, 143–146 (2000). | Article | PubMed | ISI | ChemPort |
26. Müller, D.J. et al. Observing membrane protein diffusion at subnanometer resolution. J. Mol. Biol. 327, 925–930 (2003). | Article | PubMed | ISI | ChemPort |
27. Stahlberg, H. et al. Two-dimensional crystals: a powerful approach to assess structure, function and dynamics of membrane proteins. FEBS Lett. 504, 166–172 (2001). | Article | PubMed | ISI | ChemPort |
28. Viani, M.B. et al. Probing protein-protein interactions in real time. Nat. Struct. Biol. 7, 644–647 (2000). | Article | PubMed | ISI | ChemPort |
29. Seifert, U. Dynamics of a bound membrane. Phys. Rev. E Stat. Phys. Plasmas Fluids Relat. Interdiscip. Topics 49, 3124–3127 (1994). | PubMed |
30. Sorzano, C.O. et al. XMIPP: a new generation of an open-source image processing package for electron microscopy. J. Struct. Biol. 148, 194–204 (2004). | Article | PubMed | ISI | ChemPort |

	Top

Competing interests statement:  The authors declare that they have no competing financial interests.

	Top

Natureproducts is an online service detailing information about specific products used in this article, you can view the product descriptions, request information and compare with other similar products. The products used are listed in alphabetical order. A-Z product listing DMRXA2 (Leica) field emission gun (XL-30S (Philips) J-scanner (Olympus) lithographic pattern generator system (Raith) Nanoscope-E AFM (Veeco) oxide-sharpened Si3N4 cantilevers (Olympus) See more natureproducts

	Top