Introduction
Atomic force microscopy (AFM) began as an art as well as a science, and in many ways it continues to be an art and a science. Despite the existence of new, better, faster, bigger (or smaller!) atomic force microscopes (AFMs), some of the of the fundamental challenges remain—probe tips, sample dirt, and moving beyond leading-edge results obtained with one biological system to obtain leading-edge results with a related system.
With regard to sample preparation, some methodological approaches are easily communicated. Other approaches are often nearly subconscious, such as the choice of rinsing a sample with a stream of water directed directly at the sample spot or directed above the sample spot, such that the water flows more gently over the sample spot. This can result in very different surface coverages, such that one person's samples will have many DNA molecules per square micron, whereas another person's samples will have only an occasional DNA molecule.
Different regions of the same sample can appear very different. In one early experiment, our samples of DNA with a single-stranded binding protein showed only a solid “lawn” of material. By imaging progressively farther from the center of the sample, we eventually came upon a region where the lawn ended and individual molecules of DNA coated with single-stranded binding protein were seen.
Operation of AFMs is also a combination of art and science that involves learning general rules for cantilever tuning, imaging forces, speeds, and gains and then manipulating these variables to optimize the imaging of the specific sample under the cantilever's probe tip.
Exciting breakthroughs were made in some of our earliest work, such as the time-lapse movie of fibrin polymerization [Reference Drake1] and reproducible imaging of DNA under propanol (Figure 1A) [Reference Hansma2]. Later research included time-lapse movies of DNA (Figures 1B & 2) [Reference Bezanilla3, Reference Argaman4], other research on DNA (Figure 1C) [Reference Hansma5, Reference Hansma and Laney6, Reference Golan7, Reference Sitko, Mateescu and Hansma8], the extracellular matrix [Reference Chen, Clegg and Hansma9, Reference Chen and Hansma10], synaptic vesicles [Reference Garcia11, Reference Laney12], analyses of bacterial biofilms [Reference Auerbach13, Reference Steinberger14], and pulling on spider silk [Reference Becker15, Reference Oroudjev16]. AFM methods for imaging in fluid are presented in Kindt et al.'s article in Methods in Cell Biology [Reference Kindt17], which is part of a useful volume about AFM methods in cell biology [Reference Jena, Horber, Wilson and Matsudaira18]. Several reviews by members of our lab have appeared therein [Reference Hansma and Hoh19, Reference Hansma20, Reference Hansma21].
Results
Fly eye. An undergraduate in the lab caught a fly, and Oroudjev removed the eye, mounted it, and imaged it by AFM. Individual ommatidia in the compound eye of the fly are ~20 μm wide and ~3 μm high (Figure 3). The surfaces of the ommatidia are covered with aperiodic arrays of substructures, as seen in amplitude and phase images from tapping-mode AFM (Figure 4). These substructures have been described as a cerebral cortex-like pattern of ridges in earlier AFM images [Reference Anderson and Gaimari22].
Collagenous extracellular matrix (ECM) from osteoblasts. Osteoblasts are bone-building cells that produce large quantities of extracellular collagen. Extracellular matrix from cultured osteoblasts was treated to remove calcium, as outlined in the caption of Figure 5. Collagen fibrils in the ECM can be identified by their strong banding, with 64- to 65-nm periodicities and weaker banding with 32–33 nm. The 2 × 2 μm amplitude image inset in Figure 5B shows a region almost completely covered with nearly parallel collagen fibrils.
Probe tips—still a holy grail of AFM research. Imagine that the AFM could reliably image sequence-dependent substructure on DNA at the level of the double-helix. This appears to have been done in a small series of images from our lab in the mid-1990s. Figure 6 is one of these images. This image was taken with an early prototype AFM for tapping in fluid that never became commercial. The silicon-nitride V-shaped cantilever used for the image had a “super tip” grown at the tip of the cantilever's tip by electron beam deposition (EBD) in a scanning electron microscope (SEM). The super tip started out as an unimpressive tip, giving apparent DNA widths of ~7 nm. After several images, the apparent width of the DNA changed to that seen in Figure 6, with an apparent width of only 2–3 nm, like the diameter of the DNA double helix by x-ray crystallography. After a half dozen scans, the apparent width of the dsDNA again became too broad to resolve helix turns [Reference Hansma21].
Each AFM tip is unique and changes even during use, as it picks up or loses material. Some of the best and sharpest AFM tips were the early glued diamond tips, in which a shard of diamond was glued to a tipless cantilever, using an eyebrow hair to maneuver the glue and the diamond onto the cantilever [Reference Drake1]. When carbon nanotubes were discovered, there was great hope that these would soon be standing on top of the tips on cantilevers. Twenty years later, there is still much room for improvement.
One approach that was sometimes useful for improving tips was to move to an “unimportant” area of the sample and image at high speed and high force. Although this might be expected to dull the tip, in many cases it actually improved the image quality, as determined by the width (narrowness) of small features.
Mica. The samples above, except for the fly eye, were all imaged on mica, which is a layered aluminosilicate mineral with perfect cleavage planes. The sheets of mica are held together by potassium ions, which are also present in ~100-mM concentrations in the cytoplasms of all cells. One of mica's advantages is that it is atomically flat, thus allowing one to see even bare DNA molecules as ridges above the mica surface.
Possible origin of life between mica sheets. An unexpected offshoot of biological AFM has been the hypothesis that life might have originated between mica sheets (Figure 7) [Reference Hansma23]. The spaces between mica sheets provide many adjacent chambers or “cells” in which pre-cellular life could have developed, in an environment rich in potassium ions and low in entropy. Mica also provides an endless energy source—mechanical energy or work—that might have been used to form covalent bonds, restructure polymers, and bleb daughter cells off the earliest protocells.
Mica's affinity for biomolecules provides supporting evidence that life could originate in a “Mica World,” as proposed in the mica hypothesis above. Mica's affinity for biomolecules can be seen especially in Figures 2 and 8. DNA (Figure 2) and two very different proteins—laminin (Figure 8) and RNA polymerase (Figure 2)—bind to mica. Both DNA and laminin move on the mica surface (Figures 2 and 8). The binding of DNA to mica is strengthened or weakened in the presence of different divalent inorganic cations commonly found in living systems [Reference Laney12].
New techniques in microscopy and microanalysis might be necessary for acquiring direct new evidence for the mica hypothesis. It might be possible to synthesize biomolecules between mica sheets by mechanochemistry, using the work from moving mica sheets. However, any newly synthesized molecules would be present in only sub-monolayer quantities, and the molecules would be between the mica sheets as opposed to on top of them. Detecting and analyzing these molecules would be done best with techniques that could ablate successive layers of mica and then analyze the molecules on the surfaces of the newly exposed mica sheets. Micro-analysis would be orders of magnitude too insensitive to detect these sub-monolayers of molecules. What would be needed is something more like zepto-analysis. In the meantime, our evidence that life could originate in a “Mica World” is more limited, relying on evidence such as mica's affinity for biomolecules and other supporting evidence described in [Reference Hansma23].
Acknowledgments
Thanks to Caitlin Kreutz for catching the fly whose eye's images are in Figures 3 and 4. She was one of many undergrads who were involved in much of the work presented here. Thanks to Paul Hansma for his early interest in biological AFM. Much of the early work was done on his prototype AFMs. Thanks to NSF BIO for funding the research in this paper.