Ultrahigh-speed imaging reveals nanoscopic single-molecule dynamics

Biological events (e.g., immune responses and infection by disease) originate from the interactions that take place between individual molecules at the nanoscale. Probing single-molecule dynamics is therefore a powerful method for understanding the underlying mechanisms of biology. Furthermore, because of the great complexity of biological systems and the rapid motion of single molecules, measurements with high spatial and temporal resolution are highly desired. In biological membranes, for example, the ability to track a single molecule down to the nanoscale within microseconds would enable investigations into the existence and functions of putative membrane nano-domains with molecular clarity. High-speed measurements are generally noisy, however, as a result of the limited signal-integration time as well as the introduction of electronic noise. These issues prevent close examination of nanoscopic dynamics at the single-molecule level.

To make a single molecule visible under a microscope, optical probes are used to label the molecule of interest. Among these, fluorescent tags (e.g., dyes and fluorescent proteins) and scattering tags (e.g., gold nanoparticles) are the most widely used. Fluorescent tags offer excellent contrast in non-fluorescent environments. Fluorescence suffers from photobleaching and saturation, however, limiting the amount of available signal. In contrast, scattering tags provide a stable signal with a strength that scales linearly with illumination intensity. Scattering-based imaging has therefore been an indispensable tool for high-speed and high-precision measurements. The main concern regarding the use of scattering tags is the relatively large size of the label, which may perturb native behaviors of the targeted molecules.

With the aim of elucidating single-molecule dynamics at the nanoscale, we have been working on an interferometric imaging method that allows us to track smaller scattering tags with better spatial precision and at higher speed. The concept of interferometric detection of a scattering signal was originally conceived by Vahid Sandoghdar and colleagues.¹ Since then, interferometric scattering (iSCAT) microscopy has been repeatedly demonstrated as a sensitive imaging modality for the detection of intrinsic scattered light signals from nanosized objects.^2–6 Instead of detecting the intensity of the scattering signal directly (which is very weak for small particles), iSCAT microscopy introduces a coherent reference beam that interferes with the scattering signal to enhance it. Part of the illumination is taken as the reference beam and overlapped with the signal at the detector. The resulting ‘interference cross term’ becomes the detected signal, which is much stronger than the scattering intensity. Additionally, homodyne detection of this interference pushes the imaging sensitivity to the shot-noise-limited regime, thereby circumventing the introduction of electronic noise.⁵

Our iSCAT microscope—see Figure 1(a)—uses a laser as the illumination light source. Laser light is projected onto the sample through a microscope objective, and the reflection from the coverglass that holds the sample serves as the reference beam: see Figure 1(b). The transmitted light illuminates the sample that contains the gold nanoparticles. The backscattered light from the nanoparticles and the reflected reference beam are then collected by the same objective and imaged onto an ultrahigh-speed CMOS camera. When we set the interference to be destructive, the particle appears as a dark spot on a bright background. We are then able to track the single 20nm gold nanoparticle with 2nm spatial precision at up to 500,000 frames per second.⁵ Because it is based on interferometry, iSCAT microscopy provides images at the optimal sensitivity (i.e., shot-noise limited), even at the highest acquisition rates possible.

Figure 1. (a) Schematics of interferometric scattering (iSCAT) microscopy. Laser light is projected into the sample through a beamsplitter (BS) and microscope objective (OBJ). The reflection from the coverglass that holds the sample (in this case, a gold nanoparticle) creates a reference beam. Transmitted light illuminates the sample and backscattered light from the nanoparticle is then collected, in addition to the reflected reference beam, by the same objective. The resulting image is projected through the BS and onto the CMOS camera. The inset shows an iSCAT image of a 20nm gold nanoparticle. (b) Close-up view showing the schematics of iSCAT detection at the sample.

We used our ultrahigh-speed iSCAT microscopy technique to investigate single-molecule membrane dynamics at the nanoscale.⁶ To achieve this, we attached single 20nm gold nanoparticles to lipid molecules in the reconstituted bilayer membranes and recorded their diffusion trajectories at high spatiotemporal resolution (see Figure 2). With a homogeneous membrane (i.e., one composed of a single lipid species), we observed simple Brownian motion from microseconds to seconds (see Figure 2 and video⁷). We then prepared a ternary membrane containing coexisting membrane domains of different phases (i.e., liquid-ordered, L_o, and liquid-disordered, L_d) phases. With our approach, we were able to follow single molecules as they explored both domains in a continuous manner and with unprecedented clarity (see video⁸). We also observed anomalous subdiffusion in the L_o domains at microsecond timescales.⁶ This result provides the first experimental evidence of nanoscopic substructures of the L_o phase, suggesting a new mechanism by which membrane dynamics are regulated at the nanoscale.

Figure 2. Tracking single lipid molecules in a bilayer membrane with ultrahigh-speed iSCAT microscopy. The image acquisition rate was 50,000 frames per second, corresponding to a temporal resolution of 20 microseconds. A long and precise diffusion trajectory was obtained, from which diffusion characteristics can be analyzed with great detail.

Although our technique shares the principle of interference imaging with other approaches, such as digital holography⁹ and quantitative phase imaging,¹⁰ the clever design of iSCAT microscopy makes it a convenient and robust imaging modality with high sensitivity. The common-path interferometry of iSCAT microscopy reduces the sensitivity of the signal to environmental noise. Using the weak reflection from the supporting coverglass as the local oscillator enhances the sensitivity of the device to the particles by allowing more illumination light to enter before the reference beam saturates the detector. Furthermore, by placing the object close to the focal plane of the microscope objective, the particle can be seen in the iSCAT image without digital reconstruction. Thus, the real-time acquisition of unpredictable one-time events is greatly simplified.

To understand biology at the molecular scale, techniques for visualizing the dynamics of individual molecules at high resolution are required. Our approach, based on iSCAT microscopy, has made it possible to track single lipid molecules in well-controlled artificial systems with nanometer spatial precision on microsecond timescales. In our next step, we will extend these high-precision, high-speed measurements to live cellular environments. Such measurements will extend our knowledge of cellular dynamics to a new spatiotemporal regime (i.e., nanometer and microsecond). We expect cellular background scattering to pose a challenge by making it difficult to detect signals from the particle of interest. Developing advanced image processing that is capable of differentiating between signal and background will therefore be imperative.

The authors acknowledge financial support from the Nano Program of Academia Sinica and grants from the Ministry of Science and Technology, Taiwan (102-2112-M-001-002-MY3 and 105-2112-M-001-016-MY3).