In late 1959 the visionary Richard Feynman gave his now classic talk suggesting that “there’s plenty of room at the bottom”, which forecast that in the future “we could arrange the atoms one by one the way we want them” and that high-resolution microscopes would allow a direct look at single molecules in biological samples1
. Nearly 50 years later, this prediction has spawned the fields of nanotechnology and single molecule (SM) microscopy. In the 1980’s so-called scanning probe and near-field microscopes were developed that use sharp, nanoscale tips to image, probe and manipulate individual atoms or molecules2–4
. Pioneering efforts in the late 1980’s and early 1990’s realized optical SM detection in wide-field microscopes5–8
. While at first limited to the observation of single chromophores embedded in a crystalline matrix at low temperatures, imaging of single molecules under ambient conditions soon became possible9–11
, enabling the envisioned broad applications of SM tools in biology.
Many reviews have described the unprecedented insights into complex biological processes provided by the observation and manipulation of single molecules (for a small selection of recent reviews see12–19
). Briefly, according to the ergodicity hypothesis from statistical mechanics, a sufficiently long time average (or sufficient number of observations) from a single molecule is equivalent to a standard population-averaged snapshot, suggesting that, in principle, an SM experiment contains all information of the molecular ensemble. Additionally, SM approaches: (i) reveal heterogeneity and disorder in a sample, albeit in a finite observation window (typically seconds to hours), which seem to contradict the ergodicity hypothesis but are commonplace in biological systems; (ii) afford precise localization (with nanometer accuracy) and counting of molecules (up to 105
) in spatially distributed samples such as a living cell; (iii) work at the low numbers found for most specific biopolymers (proteins, nucleic acids, polysaccharides) in a living cell (typically 1–1,000), eliminating the need for artificial enrichment; (iv) enable the quantitative measurement of the kinetics (μs to seconds) or statistics of complex biological processes without the need for a perturbing synchronization of molecules to reach a sufficient ensemble-averaged signal; (v) reveal rare and/or transient species along a reaction pathway, which are typically averaged out in ensemble measurements; (vi) enable the ultimate miniaturization and multiplexing of biological assays such as SM sequencing20
; (vii) facilitate the direct quantitative measurement of mechanical properties of single biopolymers and their assemblies, including the forces (10−2
pN) generated by biological motors; and (viii) provide a way to “just look at the thing”, as Feynman suggested1
, since one can argue that seeing SM behavior is believing. In combination, these features lead to the profound intellectual and scientific appeal of SM tools and their imminent potential to revolutionize all aspects of the biosciences including structural biology, enzymology, nanotechnology, and systems biology. However, the capabilities of existing SM techniques also have limitations, especially in the accessible measurement accuracies, time resolutions, and time windows, as posed by the weak signal and potential for loss of the observed molecule.
While many studies attest to the unique information gained from SM observation (a conservative estimate places relevant publications at currently ~2,000, with an exponentially increasing trend over the past four decades14
), two bottlenecks have impeded an even more rapid and widespread incorporation into the biological sciences. The first bottleneck derives from the perceived requirement for expansive experience and expensive equipment. The accompanying review by Ha and coworkers seeks to encourage researchers to overcome this hurdle by building their own affordable SM fluorescence microscope19
. A complementary solution is the implementation of open-access resource centers, much like existing structural biology centers15
, or other forms of collaborations with specialists. The second impediment to a broader application of SM tools in biology stems from the need to identify the most suitable technique from the toolkit and develop the corresponding assay to solve the scientific question at hand. The current review aims to provide practical “do-it-yourself” guidelines for choosing the optimal SM tool for any number of research problems. The best choice will depend on the observable of interest, which leads to an organization of the review by categories of observables. For each category, we will provide examples for successful SM assays, as well as a discussion of data analysis, limitations, and possible advances in the future. First, however, we will survey the rapidly expanding optical and force microscopy toolkit available to the SM microscopist (electrophysiology techniques as applied to single membrane-bound ion channels are beyond our scope).