Imaging, measuring and manipulating
native biomolecular systems with the
atomic force microscope

INTRODUCTION

Biological systems can only be fully understood if their structure is known. This underscores the importance of structural biology, the science investigating the structure and function of the components of living systems. In contrast to other fields of molecular biology, it is an area of research where progress depends critically on sophisticated instruments . The atomic force microscope (AFM) is one of the most powerful tools for determining the surface topography of native biomolecules at subnanometer resolution [c.f. Müller et al., 1995; Schabert et al., 1995; Shao et al., 1996]. Unlike X-ray crystallography and electron microscopy (EM), the AFM allows biomolecules to be imaged not only under physiological conditions, but also while biological processes are at work. Because of the high signal-to-noise (S/N) ratio, the detailed topological information is not restricted to crystalline specimens. Hence single biomolecules without inherent symmetry can be directly monitored in their native environment [c.f. Shao et al., 1996].

The AFM can also provide insight into the binding properties of biological systems. In order to determine the specific interaction between two kinds of molecules (e.g., avidin and biotin) in an AFM, one kind (i.e., avidin) is bound to the tip of a cantilever and the other kind (i.e., biotin) covers the surface of the sample support. The adhesion force upon separation is then a measure of the binding strength. This method allowed the intermolecular forces between individual ligand-receptor pairs [Moy et al., 1994; Florin et al., 1994], complementary DNA strands [Lee et al., 1994], cell adhesion proteoglycans [Dammer et al., 1995], and the specific antigen-antibody interaction [Dammer et al., 1996] to be determined.

PRINCIPLE

The AFM works in the same way as our fingers which touch and probe the environment when we cannot see it. By using a finger to "visualize" an object, our brain is able to deduce its topography while touching it. The resolution we can get by this method is determined by the radius of the fingertip. To achieve atomic scale resolution, a sharp stylus (radius ~1-2 nm) attached to a cantilever is used in the AFM to scan an object point by point and contouring it while a constant small force is applied to the stylus (Fig. 1a). With the AFM the role of the brain is taken over by a computer, while scanning the stylus is accomplished by a piezoelectric tube (Fig. 1b). This simple technique provides a peep into the microscopic world, and it enables us to understand how the "smallest bricks" (i.e., the biomolecules) of biological systems might function.

Figure 1. Principle of the AFM. (a) A fine stylus is mounted on a cantilever spring and scanned over the surface. At sufficiently small forces the corrugations of the scanning lines represent the surface topography of the sample. (b) The vertical deflection of the cantilever is detected by reflecting a laser beam onto a 2-segment photodiode. The photodiode signal is used to drive a servo system which controls the movement of the piezo xyz-translator. In this manner the applied force between the stylus and the sample can be kept constant within some tens of a piconewton. The imaging process can be performed in a liquid cell filled with buffer solution. This ensures that the biomolecules remain hydrated. (c) Atomic resolution of a mica surface recorded in aqueous solution. The distance between adjacent protrusions is 5.4 Å.

APPLICATIONS: From cartilage to collagen, and to the structure and function of individual proteins

With the AFM, native tissue can be directly observed without prior dehydration. This can be important as demonstrated with articular cartilage which has to be kept in physiological buffer to preserve its ultrastructure. Surface irregularities observed by scanning electron microscopy (SEM) are absent on AFM inspection. Occasionally, the cartilage surface (Fig. 2a, bottom) exhibits local discontinuities where an underlying fibrous network is distinguishable (Fig. 2a, top). Digestion of the cartilage surface with chondroitinase AC exposes this fibrous network more systematically so that the individual fibers are visualized with great clarity by AFM (Fig. 2a, inset). When imaged at higher magnification (Fig. 2b), these distinct fibers exhibit a 60nm repeat (Fig. 2b, inset), indicating that they are assembled from collagen fibrils. In addition, the micromechanical properties along the articular surface can be measured with the AFM. Hence this instrument provides new possibilities for the characterization of both surface structure and mechanical properties of freshly excised articular cartilage under physiological conditions [c.f. Jurvelin et al., 1996]. Preliminary data confirm the existence of an amorphous, non-fibrous articular surface which is expected to be vital for lubrication and wear of articular cartilage surfaces in vivo.

Figure 2. Surface and sub-surface of fresh articular cartilage from a bovine humeral head. (a) The most superficial layer, typically 200-500 nm thick, consists of acellular and non-fibrous tissue (bottom). Occasionally, this exhibits local discontinuities where the underlying network of collagen fibrils oriented parallel to the surface becomes visible (top). Treatment with chondroitinase AC removes the majority of the articular surface and exposes the superficial collagen fibrils (inset). (b) Higher magnification view of the collagen fibrils which exhibit a mean diameter of ~35nm. Tangential fibrils reveal a characteristic periodic banding pattern of 60±5nm (inset).

Figure 3. AFM topographs of purple membrane from Halobacterium salinarium. Purple membrane consists of 25% lipids and 75% bacteriorhodopsin. This light driven proton pump comprises 7 transmembrane a-helices which surround the photoactive retinal. (a) Imaged at forces of about 3x10 -10 N two of the three loops connecting the a-helices are visible on the cytoplasmic surface (inset). (b) When the applied force is reduced during imaging (from 3x10 -10 N at the beginning (bottom) to 1x10-10 N at the end (top) of the scan), the proteins undergo a conformational change. (c) The most prominent loop connecting the a-helices E and F is imaged at 1x10 -10 N, but is bent toward the membrane surface at higher forces.

As a result of their high S/N ratio unprocessed AFM images provide atomic detail of solids. In contrast, biomolecules such as proteins which are designed to undergo conformational changes and form flexible supramolecular assemblies , are mechanically "soft", i.e., they are best compared with a "sponge". Hence their surface topography cannot be probed at atomic detail by AFM. Nevertheless, state-of-the-art specimen preparation and instrumentation now allow the surface topography of native proteins to be imaged at subnanometer resolution. As documented in Fig. 3 [c.f. Müller et al., 1995; 1996a], by scanning the cytoplasmic surface of the purple membrane of the Halobacterium salinarium with a very low cantilever force (i.e., ~10-10 Newton), the most prominent loop connecting the transmembrane a-helices E and F of the bacteriorhodopsin molecule can be imaged in its extended conformation (Fig. 3b, top; Fig. 3c). As the cantilever force is increased to ~3x10-10 Newton, this prominent loop is mechanically bent toward the membrane surface (Fig. 3a and 3b, bottom; Fig. 3c). The energy required for this reversible structural change is approximately 6kJ/mol, much too small to disrupt single chemical or electrostatic bonds.

Figure 4. Conformational change of the hexagonally packed intermediate (HPI) layer of Deinoccocus radiodurans. The HPI layer consists of units which display a large central pore and are assembled from six protomers. The function of the surface layer is not understood, but it is thought to protect the cell from hostile factors of the environment, and might also be responsible for the uptake and release of nutrients and cellular signals. (a) Single pores of the inner surface occur in "open" (circled) and "closed" (boxed) conformations. (b) Imaging the same surface area after 5 min, some of the pores have changed their conformation. This conformational change is fully reversible and can be observed over hours. (c) Averages of open and closed pores as depicted in (a) and (b) together with a calculated difference image.

As illustrated in Fig. 4 [c.f. Müller et al., 1996b], at nanometer scale resolution AFM images of the cytoplasmic surface of the hexagonally packed intermediate (HPI) layer of the bacterium Deinoccocus radiodurans reveal protruding protein cores. Being formed from six protomers these are connected by emanating arms and contain a central pore. The pores exhibit two conformations: a "closed" (Fig. 4a, boxed pores; Fig. 4c) with, and an "open" (Fig. 4a, circled pores; Fig. 4d) without a central plug. Most exciting, over time a stochastic opening and closing of individual pores is monitored with the AFM (Fig. 4a and 4b; compare corresponding boxed and circled pores).

The extracellular and periplasmic surface topographies of 2-D crystalline lipid-protein arrays of the E. coli outer membrane protein OmpF have been determined with the AFM (Fig. 5a, left), and the accuracy of the corresponding surface contours has been assessed by comparison with the atomic structure of the OmpF trimer (Fig. 5a, right). The atomic model of the 2-D lipid-OmpF crystals which has been built so as to optimally fit the surface contours measured with the AFM, illustrates how membrane proteins interact with the lipid bilayer (Fig. 5b).

Identification by AFM of individual amino acid residues at protein surfaces appears now to be feasible. However, this requires a better understanding of the interactions between stylus and biomolecule , and further developments of AFM techniques. For high resolution imaging both the radius and the geometry of the stylus, and the sensitivity of the deflection detector are critical and require improvement. In addition, biological specimens must be immobilized onto ultraflat supports to decrease their lateral flexibility [c.f. Karrasch et al., 1993]. This can be achieved by chemisorption of the sample to activated substrates , or by its physisorption to atomically flat layered crystals . Our results emphasize that AFM is an useful technique in structural biology. Today, AFM images provide information on the surface structure of biomolecular systems which is complementary to other established techniques such as light and electron microscopy, nuclear magnetic resonance and X-ray crystallography. The advantage of directly observing biomolecular systems in their native environment opens the exciting possibility to analyze their structural and functional properties at the submolecular level .

Figure 5. (a) High resolution topograph of the perisplasmic surface of 2-D OmpF porin crystals. The average shown was calculated from 25 translationally aligned subframes. To allow comparison, the electron density based on the X-ray structure of the OmpF trimer was rendered at a lateral resolution of 15 Å. Zebra-like contours (dark and light blue) on the overlaid transparent AFM topograph mark zones of identical altitude, with a height difference between contours of 1 Å. (b) Atomic model of protein-protein and lipid-protein interactions based on the X-ray structure of the porin trimer.

REFERENCES

Dammer, U., Hegner, M., Anselmetti, D., Wagner, P., Dreier, M., Huber, W. & Güntherodt, H.-J. (1995). Specific antigen/antibody reactions measured by force microscopy. Biophysical Journal, in press.

Dammer, U., Popescu, O., Wagner, P., Anselmetti, D., Güntherodt, H.-J. & Misevic, G.N. (1995). Binding strength between cell adhesion proteoglycans measured by atomic force microscopy. Science 267, 1173-1175.

Florin, E.-L., Vincent, T.M. & Gaub, H.E. (1994). Adhesion forces between individual ligand-receptor pairs. Science 264, 415-417.

Jurvelin, J.S., Müller, D.J., Wong, M., Studer, D., Engel, A. & Hunziker, E.B. (1996). Surface and sub-surface morphology of bovine humeral articular cartilage as assessed by atomic force- and transmission electron microscopy. Journal of Structural Biology 117, 45-54.

Karrasch, S., Dolder, M., Hoh, J., Schabert, F., Ramsden, J. & Engel, A. (1993). Covalent binding of biological samples to solid supports for scanning probe microscopy in buffer solution. Biophysical Journal 65, 2437-2446.

Lee, G.U., Chrisey, L.A. & Colton, R.J. (1994). Direct measurements of the forces between complementary strands of DNA. Science 266, 771-773.

Moy, V.T., Florin, E.L. & Gaub, H.E. (1994). Intermolecular forces and energies between ligands and receptors. Science 266, 257-259.

Müller, D.J., Büldt, G. & Engel, A. (1995). Force-induced conformational change of bacteriorhodopsin. Journal of Molecular Biology 249, 239-243.

Müller, D.J., Schoenenberger, C.-A., Büldt, G. & Engel, A. (1996a). Immuno-atomic force microscopy of purple membrane. Biophysical Journal 70 , 1796-1802.

Müller, D.J., Baumeister, W. & Engel, A. (1996b). Conformational change of the hexagonally packed intermediate layer imaged by atomic force microscopy. Journal of Bacteriology 178, 3025-3030.

Schabert, F.A., Henn, C. & Engel, A. (1995). Native Escherichia coli OmpF porin surfaces probed by atomic force microscopy. Science 268, 92-94.

Shao, Z., Mou, J., Czajkowsky, D.M., Yang, J. & Yuan, J.-Y. (1996). Biological atomic force microscopy: What is achieved and what is needed. Advances in Physics 45, 1-86.