Ultrasonic motors can avoid sample vibration and provide stable positioning that allows for superresolution imaging of fluorescence during protein binding.
Jonas Fischer and Jürgen Schmied, Massive Photonics, and Cameron Hughes and Thomas Bocher, Physik Instrumente (PI)
During the last two decades, research in the field of superresolution microscopy has been focused on incorporating different techniques to increase the spatial resolution in collected images. As these techniques have matured, this focus in microscope development has shifted toward the goal of imaging a maximum number of targets within proteins and cells, without a loss in resolution, requiring optimal sample stability. To this end, the accuracy and reliability of the microscope positioning stages is a key factor to attain this stability as well as achieve high throughput and cost efficiency.
Multiplex image acquisition
A subset of superresolution techniques that routinely deliver unparalleled resolution, high-labeling specificity, and advanced multiplexing capabilities are single-molecule localization microscopy (SMLM) techniques. These methods have been used to examine the complex structure of biological systems at the molecular level.
Figure 1. Microtubules (magenta),
TOM20 (cyan), and PMP70 (yellow)
stained with the MASSIVE-AB 3-PLEX
kit and sequentially imaged using
Cy3b as a dye. Scale bar: 5 μm. Courtesy
of Massive Photonics.
This focus in microscope development has shifted toward the goal of imaging a maximum number of targets within proteins and cells, without a loss in resolution, requiring optimal sample stability.
SMLM techniques acquire a signal from only a stochastically chosen subset of targets at any one time. From each diffraction-limited spot, the position of the corresponding emitter can be localized with high precision and a standard deviation much smaller than the blurred spot. The resolution is tied to the number of photons collected from each event. To localize a single emitter, the emitter density must be low enough that only one emitter is detectable in a single diffraction-limited spot.
By separating the detection of subsets temporally, all the emitter positions can be calculated and a superresolved image of the target structure, such as DNA or viral proteins, can be reconstructed. Over the course of a single measurement, the same target is detected multiple times. The stochastically switching fluorescence signal can be achieved either photo-physically, by means of switching molecules between a fluorescent “on” and nonfluorescent “off” state, or by other biophysical means as described in the following section.
DNA-based points accumulation for imaging in nanoscale topography (DNA-PAINT) is a specific approach in which the stochastic activation of fluorescence is achieved by using binders labeled with a short DNA strand. A dye-labeled oligonucleotide, called an imager, binds transiently to the strand attached to the target (docking strand). The imager strands are not detectable when they are freely diffusing in solution; but when binding to the target sequence, they dwell long enough so that more photons can be collected from a single binding event.
Imager strands are freely diffusing in the solution around the sample, acting as a reservoir, exchanging imagers with the sample and creating a homogeneous background (Figure 1). In this process, photobleached dyes are continuously replaced, allowing for longer imaging times without damaged dye molecules affecting the sampling rate of each target.
Figure 2. An illustration of sequential multiplexing. Binders attach to their respective targets and have unique DNA-based points
accumulation for imaging in nanoscale topography (DNA-PAINT) binding sites. In each step, only the imager strand with the sequence,
complementary against one target, is added to the sample. Then, the data acquisition is conducted, and the imager is washed out
before the next imager is added until all targets are imaged. Courtesy of Massive Photonics.
To observe multiple targets in the same sample, multiplexing strategies can be employed. Though using different dyes for each target is the approach most widely used in research laboratories, the approach is limited by the number of dyes whose emission spectra can be separated from each other.
By contrast, the unique properties of DNA hybridization allow for theoretically unlimited multiplexing (Figure 2). By using orthogonal docking sequences, each imager binds only to its complementary strand. The transient interactions also allow for washing imagers out of the sample that can be replaced with imagers that bind to a different docking strand species. Being able to wash out the imager strands allows for the same fluorescent dye to be used with each imager, eliminating the need for further post-processing steps used in other modalities that correct for chromatic aberrations and differences in resolution.
Figure 3a. A schematic diagram of an ultrasonic-piezomotor. The piezoelectric transducer
(actuator) is preloaded against the runner. Electrical excitement of the transducer causes
oscillation. This oscillation is converted to motion, which is then transmitted to the runner
using a coupling element. The position of the runner is recorded by a stationary sensor
(encoder), which counts the periods of a grating attached to the runner. Courtesy of PI.
Figure 3b. The ultrasonic drive principle allows for the design of highly compact microscope
stages featuring a wide dynamic velocity range of six orders of magnitude (100 nm/s
to 120 mm/s) and excellent long-term stability. Courtesy of PI.
It is thus possible to observe the organization and interaction of an unparalleled number of molecular species at high resolution.
Interference during measurement
Measuring multiple targets at different positions requires that the same position be visited repeatedly during image acquisition, and that for the long duration of the measurements, the sample remains in focus. When measuring the 3D position of targets, relatively small changes in the xyz position of the sample can result in the diffraction-limited spots being out of focus. For 3D applications, such as live-cell imaging, the lateral position is usually encoded in the diffraction-limited spots, most commonly either through the introduction of a cylindrical lens or biplane imaging. If the sample position changes during imaging, the lateral position that is calculated diverges from the actual position.
Several factors affect the sample stability during measurement and lead to the position of the sample changing over time, which can result in the misinterpretation of images. They can be broadly separated into factors, causing fast and slow drift. High-frequency vibrations in the microscope setup distort the diffraction-limited spots within a single frame of the measurement, resulting in reduced resolution.
Thermal drift occurs when a difference in temperature leads to parts of the microscope setup expanding and contracting, changing the position of the sample over time. Vibrations from the setup, the surrounding environment, and stresses within the sample holders also cause slow drift within the time frame of the measurement.
To track the drift of the sample fiducials, either fluorescent beads or gold particles can be employed to track the sample movement over time. In post-processing, their positional information can be used in conjunction with other drift correction algorithms to correct slow-moving drift if it is not overly excessive or occurs within the acquisition of a single frame of the measurement.
Other sources of positional errors occur between measurements. For sequential multiplexing to work, imaging solutions containing different imagers must be exchanged before measuring the next target. Moving the sample in that time either accidentally or deliberately can be detrimental to the ongoing experiment, if the stage is incapable of detecting the positional change and correcting it before the next measurement.
In preparation of a typical measurement, the microscope stage is moved to the desired positions and their xyz coordinates are recorded. For each imaging round, the stage moves between each position and enters a closed loop mode during the acquisition of each position. After the acquisition at each position is completed, the imaging solution in the sample is exchanged and the stage moves to each position again, continuing acquisitions for the next target. This process is repeated until all targets are precisely imaged at each position.
For this repetitive procedure, a very good bidirectional reproducibility of the xy-sample stage is a must. In this case, ultrasonic-piezomotor stages offer a significantly better bidirectional reproducibility of ±600 nm versus stepper stages with ≥±1 µm.
Because these measurements can take between minutes and several days, depending on their experimental complexity and desired target number, they require the best positioning solutions available.
Ultrasonic motors
Ultrasonic microscope stages provide an ideal platform for positioning samples in a way that is compatible with the demanding conditions required for superresolution microscopy with DNA-PAINT. Such stages employ an ultrasonic-piezomotor capable of directly driving the moving stage platform (Figures 3a,b). A piezoelectric transducer, which creates motion in nanometer increments at a frequency >100 kHz, is preloaded against a runner using a coupling element. When comparing the ultrasonic drive principle with conventional stepper motors or electromagnetic linear motors, a few things stand out. Stepper motors cannot provide the high dynamic range of the ultrasonic motor and induce vibrations due to their stepping nature. Electromagnetic linear motors provide favorable dynamic properties, but their coils must be powered constantly to hold a position.
Due to the high stability, responsiveness, resolution, and favorable repeatability of ultrasonic motor microscope stages, they provide benefits to all fluorescence microscopy techniques used in live-cell imaging. Techniques that can potentially benefit include superresolution microscopy, wide-field fluorescence microscopy, structured illumination, total internal reflection fluorescence microscopy, and digital fluorescence slide scanning.
Electrical excitement of the piezoelectric actuator at its resonance frequency causes oscillation. Due to the preload, the actuator oscillation is converted into continuous feed motion by the coupling element, which moves the runner. The preload also causes the drive to self-lock when the stage is not energized, providing reliable mechanical stability to the stage — there are no lead screws that heat up during motion and no lubricants to cause drift. At rest, there is also no servo dither to deal with, as can be the case with electromagnetic linear motors.
In preparation of a typical measurement, the microscope stage is moved to the desired positions and their xyz coordinates are recorded.
The velocity of the motion can be adjusted by modifying the amplitude of the excitation, and therefore the amount of power transferred to the runner. Changes in position of the stage are detected accurately by an incremental, or in some cases, an absolute-measuring linear encoder. The number of counts recorded by the encoder is proportional to the distance traveled.
This ultrasonic motor design is compact and allows integration into a low-profile stage that integrates easily into a commercial or home-built microscope; this stands in contrast to relatively bulky stepper motor/leadscrew assemblies. By adjusting the amplitude of the oscillation, it is possible to tune the velocity over more than six orders of magnitude (120 to 100 nm/s). The driving method of the motor generates minimal vibrations, noise, and heat, providing a suitable working environment for the sample under investigation as well as the operator. Steps down to 100 nm in size are possible, and the stages can be easily combined with piezo scanner systems for movements down to the single nanometer range or below. Finally, by nature of the self-locking design inherent to ultrasonic-piezomotors and through appropriate selection of materials for the stage, long-term drift values of <50 nm are achievable (Figures 4a,b).
Figure 4. Representative data from an ultrasonic positioning stage showing low long-term
stability, better than 50 nm (a) and a series of 1-μm steps (b). Courtesy of PI.
A number of recent developments in microscope staging further improve the suitability for demanding applications, such as superresolution microscopy. To obtain higher motion accuracy and fast response over a large velocity range, adaptive parameter control — which estimates variations in a dynamic sample — is used with state-of-the-art ultrasonic microscope stages. Since piezoelectric materials and ultrasonic motors exhibit nonlinear behavior, small steps and low-velocity operations require different control parameters, compared with long-distance moves at high velocity. Using one optimized set of control parameters, it is possible to reduce drive offset to the set point and settling time for the stage. For nanometer-scale incremental motion as well as noiseless operation, a dual-frequency drive mode has been developed that drives the piezoelectric transducer at two separate frequencies. It helps to control and influence the transition between dynamic and static friction to improve motor performance.
DNA-PAINT allows for multiplexed imaging that enables complex analysis of cellular structures and interactions. The technique is uniquely supported by ultrasonic sample stages, because they allow precise motion with small step size, minimal noise, heat, and drift.
Meet the authors
Jonas Fischer studied physics with a specialization in biophysics at Ludwig Maximilian University in Munich, Germany, and is a scientist at Massive Photonics specializing in technology development for superresolution microscopy; email: f[email protected].
Cameron Hughes received his Ph.D. in physics at Caltech and, after stints in color analysis and cryogenics industries, is a product manager for Piezo Motors and Systems at Physik Instrumente (PI) GmbH & Co. KG; email: [email protected].
Jürgen Schmied holds a degree in biophysics from the University of Bayreuth, Germany, and received his Ph.D. in biophysical chemistry from the University of Technology in Brunswick, Germany. After finishing his Ph.D., he cofounded GATTAquant in 2014, focusing on calibration tools for superresolution microscopy. In 2019, he then cofounded Massive Photonics as CEO, focusing on commercializing the DNA-PAINT technology; email: [email protected].
Thomas Bocher served for ten years as product manager and business unit manager at Carl Zeiss Microscopy and seven years at Bruker BioSpin as business unit manager in compact magnetic resonance. Since 2018, he has been head of segment marketing for microscopy and life sciences at Physik Instrumente (PI) GmbH & Co. KG; email: [email protected].