Key Concepts:

Figure 1. Atomic force microscopy layout and operation. (a) General layout of an AFM setup in a sample scanner setup configuration. The deflection of a cantilever with a sharp tip at its end is monitored by means of a focused leaser beam that reflects from the cantilever on to a quadrant photodiode. A piezoelectric scanner moves the sample in the three dimensions. In this schematic representation, at time t1 the tip is positioned next to a spherical object, whereas at t2 the object is positioned below the tip, resulting in the deflection of the cantilever that will be detected as a light imbalance between the top and bottom photodiode segments. (b) Scanning electron micrograph of an AFM cantilever. The inset shows a zoom of the tip. Both scale bars indicate 2 m. Image provided by Maarten van Es and Tjerk Oosterkamp. (c) Standard modes of operation of AFM as an imaging device. Time points t1 and t2 correspond to the sample positions as in (a). The curves represent the response of the AFM feedback electronics as a function of time in (i) constant force mode or contact mode and in (ii) tapping or intermittent contact mode. The error deflection signal of the cantilever is represented by a continuous line, whereas the feedback output corresponding to the vertical position of the piezo is indicated by a dotted line. Further details on these modes are provided in the main text. (d) Example of topographic (i) and phase (ii) images obtained on a polymeric material with AFM in IC mode. Image dimensions are 2×2 m². Reprinted from Porter et al. (2000). © Nature Publishing Group. (e) Tip convolution effects. The observed width of a nondeformable topographic feature of the sample, represented by the dotted line, depends on both the size of the feature and the geometry of the tip (i). Large objects will obscure adjacent features, rendering deep cavities non accessible (ii). Tips that combine a small end radius with a high aspect ratio such as a nanotube can significantly relieve tip convolution effects and improve the lateral resolution. (f) Schematic representations of force versus distance z plots (that is, F–z curves) obtained on either a stiff (that is, nondeformable) or soft (that is, deformable) sample (i) a soft sample exhibiting a single tip-sample unbinding event (ii) and a stiff sample where multiple unfolding events appear (iii). Contact and noncontact regions are highlighted in grey and white, respectively. Black/grey lines correspond to forces recorded during tip-sample approach/retraction. Note that stiff samples are identified by the appearance of a flat tilted line in the contact region of an F–z curve, whereas soft samples are identified when a curved line appears in the same region.

Figure 2. Illustrative examples of AFM studies on single biomolecules and supported membranes. (a) Visualisation of the transition of a protein hR/M–DNA complex from associated to a dissociated conformation in liquid. Topographic AFM images were obtained in IC mode and rendered in 3D. hR/M is a human DNA repair protein consisting of a globular DNA-binding domain from which two 50 nm long coiled coils protrude. When these coiled-coils self-associate, they inhibit the protein interaction required for DNA tethering and subsequent repair. The first image shows the globular domain of a single hR/M bound to DNA, and its two-coiled coils arranged in a parallel (nonassociated) conformation in the top-left part (i). The second image shows a globular domain that is dissociated from the DNA, and how the two coiled-coils form a loop, which corresponds to the inhibitory conformation. Scale bars indicate 25 nm. Reprinted from Moreno-Herrero et al. (2005). © Nature Publishing Group. (b) Indirect detection of kinase activity with AFM used as a chemical force sensor. An Au substrata was coated with the sequence-specific peptide (i) that is specifically phosphorylated by protein kinase CK2. An Au tip was functionalised with the antiphosphorylated peptide antibody (i). Right panel shows typical F–z curves representing interactions of the functionalized tip with substrata coated with either the unphosphorylated (top) or phosphorylated peptide (bottom). Note that no rupture detachment forces are observed at the top, whereas multiple rupture forces appear at the bottom F–z curve. Rupture forces were quantified as multiples of 120±20 pN, and attributed to the dissociation of a simple binding event between the phosphorylated peptide and the antibody binding region. Reprinted from Wilner et al. (2008). © Wiley–Blackwell. (c) Height images of a helical fibril of a lithostathine protein in liquid (i). Profile delineated in (i) by the black line (ii). Lithostathine is overexpressed in preclinical stages of Alzheimer's disease and is present in lesions associated with this disease. As other amyloid peptides, two lithostathine protofibrils can associate to form an helical fibril as illustrated in this image. The inset corresponds to the filtered image, highlighting the helical nature of the fibre. This image illustrates the feasibility of observing formation of fibrillar structures of proteins, a critical process in neurodegenerative diseases. Scale bar, 30 nm. Reprinted from Milhiet et al. (2010) (available under Creative Commons Attribution-Noncommercial licence) © PloS One 2010. (d) Unfolding experiment of a multidomain protein by AFM. Schematic representation of the experiment (i), showing a multimodular protein tethered between the surface of an AFM tip and the surface of a flat substrata attached to the piezo positioner. (ii) Retraction curve obtained in an F–z curve recorded on a recombinant poly I27 consisting of eight repeated domains. The curve shows a saw-tooth pattern with 8 peaks with a periodicity of 25–28 nm, corresponding to the unfolding of each domain. When a single domain begins to be stretched, the force increases owing to the entropic elastic resistance of the protein domain. At greater extensions, the probability of unfolding becomes very high and, when the domain unfolds, the force suddenly relaxes back to its resting position. Each unfolding event was fitted to the WLC model. (iii) Frequency histogram of the unfolding forces, revealing a peak ~200 pN. Reprinted from Fisher et al. (1999). © Wiley-Blackwell. (e) Height image of a native photosynthetic bacterial membrane adapted to high light. Note the apperance of either small or large ring-structures, which correspond to light-harvesting complexes 2 (LH2) rings and core complexes, respectively. This study revealed that the ratio of LH2 rings to core complex increased from ~3.5 to ~7 in bacteria subjected from high to low light conditions. Bar, 10 nm; colour scale, 3.1 nm. Images were obtained in contact mode in liquid with a low loading force (~100 pN). Reprinted with permission of AAAS from Scheuring and Sturgis (2005). http://www.sciencemag.org.

Figure 3. Illustrative examples of AFM studies on single viruses and prokaryotic cells. (a) Topography of a single 29 bacteriophage, a type of virus that infects bacteria. The image was obtained on 29 deposited on a mica surface in liquid using frequency-modulation AFM, which is a recent modification of the standard IC mode that can detect forces down to 20 pN. The image was rendered 3D. The three main domains of the viral particle can be easily identified, including the prolate head, the collar region and the tail (marked with the arrow). In addition, fine structural features such as the corrugation of the capsid due to the capsomeric subunit arrangement are also clearly seen. The capside diameter is 25 nm. Reprinted from Martinez-Martin et al. (2012) (available under Creative Commons Attribution-Noncommercial licence). © PloS One 2010. (b) Low- (i) and high-resolution (ii) images of a single Saccharomyces cerevisiae yeast cell in liquid obtained with AFM in contact mode. Images display the error deflection signal, which highlights sudden changes in the topography as described in the main text. Note the presence of a bud scar in the left image. Yeast cells were trapped into a porous polymer membrane for immobilisation purposes. Reprinted from Dupres et al. (2009). © Nature Publishing Group. (c) The flocculation or asexual aggregation of Saccharomyces carlsbergensis yeast cells is mediated through specific lectin–carbohydrate interaction forces. To probe these forces, F–z curves were acquired between single S. carlsbergensis yeast cells and tips functionalised with carbohydrate molecules (i). Representative retraction curves (ii) from these F–z curves reveal single or multiple unbinding events, which are interpreted as the unbinding of single or multiple receptor–ligands, respectively. Curves were offset with respect to the zero force. Reprinted from Dague et al. (2007). © Wiley-Blackwell. (d) Visualisation of a spore chain of Streptomyces coelicolor bacteria obtained with AFM in IC mode. The image displays phase data, and shows an instant of the division of these bacteria during sporulation. Reprinted from Wright et al. (2010). © Nature Publishing Group.

Figure 4. AFM imaging and analysis of single eukaryotic cells and tissue sections. (a) Image of the aggregation of the polysaccharide K-carrageenan from algae, resulting in the formation of a network. A solution of carrageenan was deposited on mica, air-dried and imaged in contact mode. Image size is 1×1 m. Reprinted from Morris et al. (1997). © Elsevier. (b) Visualisation of cell morphological and mechanical changes upon perturbation of the actin cytoskeleton with the drug jasplakinolide. AFM deflection error images obtained in contact mode in liquid on a single 3T3 fibroblast before (i) and after (ii) addition of the drug. After drug treatment, the active regions start to retract and the Young's modulus decreases (iv) compared to the reference elasticity map of untreated cells (iii). Reproduced from Rotsch and Radmacher (2000). (c) Detection and analysis of the beating of a single cardiomyocyte. An AFM tip was brought into gentle contact with a single cardiomyocyte with a loading force of ~100 pN; the position of the cantilever was held for several seconds, and the corresponding force F was recorded as a function of time t (F versus t). The contraction of the cell appears as peaks in the F versus t curve. These peaks were analysed in terms of their height, full width at half maximum and reciprocal of beat-to-beat peak separation to characterise the force, duration and frequency of cardiomyocyte beats, respectively. Conducting these measurements on cardiomyocytes derived from subjects with dilated cardiomyopathy showed decreased force and cellular elasticity compared to cells from healty donors. These data reveal that AFM could be used to either screen for cardiac-active pharmacological agents, or as a platform for studying cardiomyocyte biology. Reprinted from Liu et al. (2012) (available under Creative Commons Attribution-Noncommercial licence). © PloS One 2012. (d) Detection of mechanical reinforcement in single focal adhesion (FA) precursors. Flat-ended cylindrical AFM tips with ~1 m² cross-section were coated with the integrin-specific binding domain RGD, brought into contact with the surface of a fibroblast, and held for ~30 s to enable the formation of FA precursors. The Young's modulus E, which is indicative of local resistance to deformation, was measured afterwards and found to be two-fold larger than that measured with bare or tips coated with control peptides. These measurements revealed that mechanical loading applied specifically to integrin-rich focal adhesion precursors is sufficient to orchestrate a rapid mechanoresponse that elicit strengthening or reinforcement of cell adhesions. Reprinted from Acerbi et al. (2012) (available under Creative Commons Attribution-Noncommercial licence). © PloS One 2012. (e) Real-time visualisation of the localisation of vascular endothelial growth factor (VEGF) receptor-2/Flk-1 in the surface of a single endothelial cell obtained with AFM used as a molecular recognition sensor. For this purpose, an AFM tip was functionalised with antiFlk-1 antibody, a receptor for VEGF. Recognition adhesion force maps before (i), 10 min (ii) 45 min and (iii) after starting competitive inhibition by adding antiFlk-1 antibodies to the culture medium. The micrometer-sized brigher spots in (iii) are interpreted as clusters of VEGF receptors. Reproduced from Almqvist et al. (2004) with permission of the Biophysical Society. (f) High-resolution image of the surface of wild-type and mutant corneal tissue from a Drosophila melanogaster insect. The cornea of Drosophila and other insects include lenses that are covered by an ordered array of nipples. AFM was used to obtain a 3D rendered high-resolution topographic image of a nipple array in wild-type corneal tissue (i) using IC mode in air. In contrast, AFM imaging of a cornea from a Drosophila mutant overexpressing Wg (ii) revealed a dramatic loss of nipples and a marked increase in the spacing between them, thereby implicating the Wg signalling pathway in the development of normal corneal tissue. Reprinted from Kryuchkov et al. (2011) (available under Creative Commons Attribution-Noncommercial licence). © PloS One 2011. (g) Stiffness maps of normal and fibrotic mouse lung parenchymal tissue. Colour bar indicates shear modulus G, which is proportional to E for elastic materials. AFM F–z curves were acquired in a 16×16 sample grid separated by 5 m spatially covering 80×80 m² area. G at each point on the grid was calculated from fitting F–z data with a spherical contact elastic model. Reprinted from Tschumperlin et al. (2009) Journal of Biomechanics. © Elsevier.

Figure 5. Illustrative examples of current trends in AFM applications in biological sciences and biomedical research. (a) Cantilever arrays are being used as nanosensors as part of diagnostic devices. Low magnification (60×) Scanning electron microscopy micrograph of an array of silicon cantilevers (i) Like the AFM probes, these cantilever arrays are microfabricated using semiconductor lithographic techniques. Reprinted from Wilson et al. (2010) (available under Creative Commons Attribution-Noncommercial licence). Because cantilevers can be coated with molecules such as oligonucleotides, antibodies or other proteins capable of binding to specific biomolecules, devices based on these cantilever arrays are being developed as nanosensors for diagnostic applications. The physical properties of the cantilevers change as a result of the binding event, and these changes can be monitored using suitable electronic readout strategies, thereby providing not only information about the presence and the absence but also the concentration of different disease-related molecules. This principle of operation is represented in (ii). Using a prototype based on an array of cantilevers and an optimised cantilever geometry it was possible to detect the prostate cancer marker PSA in a fluid sample at a concentration one order of magnitude lower than current ELISA-based diagnostic tools (iii). Reprinted from Kodera et al. (2010). © Nature Publishing Group. (b) Recent advances have improved the time resolution of AFM as an imaging instrument from minutes to tens of milliseconds per frame, enabling watching single biomolecules at work in real-time. As a proof-of-principle of the possibilities of high-speed AFM imaging, a single myosin V molecular motor could be observed translocating along an actin filament by recording frames at 147 ms per frame. Vertical dashed lines indicate the same positions of the actin filament across frames. Images were obtained in liquid. This succession of images demonstrates that probing structure and dynamics of single biomolecules is now feasible. Scan area, 130×65 nm²; scale bar, 30 nm. Reprinted from Wu et al. (2001).