Ex) Article Title, Author, Keywords
Current Optics
and Photonics
Ex) Article Title, Author, Keywords
Curr. Opt. Photon. 2022; 6(6): 550-564
Published online December 25, 2022 https://doi.org/10.3807/COPP.2022.6.6.550
Copyright © Optical Society of Korea.
Sangjun Park1,2, Cheol Hong Min1,2, Seokyoung Han3, Eunjin Choi1,2, Kyung-Ok Cho2,4, Hyun-Jong Jang2,5, Moonseok Kim1,2
Corresponding author: *moonseok@catholic.ac.kr, ORCID 0000-0003-3100-8980
†These authors contributed equally to this paper.
This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http://creativecommons.org/licenses/by-nc/4.0/) which permits unrestricted non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited.
Optical microscopy is a useful tool for study in the biological sciences. With an optical microscope, we can observe the micro world of life such as tissues, cells, and proteins. A fluorescent dye or a fluorescent protein provides an opportunity to mark a specific target in the crowd of biological samples, so that an image of a specific target can be observed by an optical microscope. The optical microscope, however, is constrained in resolution due to diffraction limit. Super-resolution microscopy made a breakthrough with this diffraction limit. Using a super-resolution microscope, many biomolecules are observed beyond the diffraction limit in cells. In the case of volumetric imaging, the super-resolution techniques are only applied to a limited area due to long imaging time, multiple scattering of photons, and sample-induced aberration in deep tissue. In this article, we review recent advances in superresolution microscopy for volumetric imaging. The super-resolution techniques have been integrated with various modalities, such as a line-scan confocal microscope, a spinning disk confocal microscope, a light sheet microscope, and point spread function engineering. Super-resolution microscopy combined with adaptive optics by compensating for wave distortions is a promising method for deep tissue imaging and biomedical applications.
Keywords: Adaptive optics, Super-resolution microscope, Volumetric imaging, Wavefront shaping
OCIS codes: (180.2520) Fluorescence microscopy; (180.6900) Three-dimensional microscopy; (220.1000) Aberration compensation; (220.1080) Active or adaptive optics
Optical microscopy has long been useful to study biological sciences. With an optical microscope, we take an image and get information on the micro world. Among many proteins, DNA, and RNA in a biological sample, fluorescent dye or a fluorescent protein provides an opportunity to mark a specific target. The optical resolution, however, is limited by diffraction. The diffraction limit is known as the Abbe limit as in the following equation [1]:
where
As shown in Fig. 1, mammalian cells and
The techniques of the super-resolution microscope are presented in several categories theoretically and experimentally. The major techniques, for example, are stimulated emission depletion (STED) microscope [5–12], structured illumination microscope (SIM) [13–22], photo-activated localization microscope (PALM) [23–28], stochastic optical reconstruction microscope (STORM) [29–37], fluorescence photo-activated localization microscope (FPALM) [38–40], interferometric photoactivated localization microscopy (iPALM) [41, 42], 4Pi single marker switching nanoscopy (4Pi-SMSN) [43], reversible saturable optical linear fluorescence transitions microscope (RESOLFT) [44–48], super-resolution optical fluctuation imaging (SOFI) [49–54], DNA point accumulation for imaging in nanoscale topography (DNA-PAINT) [55–61], and FRET-based DNA-PAINT (FRET-PAINT) [62–65].
In this article, we review the methods of super-resolution microscopy combined with adaptive optics for volumetric imaging. In volumetric imaging, the super-resolution microscopes have been applied to a limited area due to long imaging time, multiple scattering, and sample-induced aberration in deep tissue. To break through the technical barriers, the super-resolution microscopes have been integrated with line-scan confocal microscopy, spinning disk confocal microscopy, light sheet microscopy, or point spread function engineering. The super-resolution microscopy incorporated with adaptive optics has been exploited for volumetric imaging by compensating for the wave distortions in biological tissue. As a result of the improvement, super-resolution microscopy with adaptive optical techniques can be advantageous for deep tissue imaging and biomedical applications. Among several methods of super-resolution imaging, we extensively discuss STED, SIM, and single-molecule localization microscopy with adaptive optics in this review.
The super-resolution microscope offers the advantage of breaking the diffraction limit and obtaining images of biomolecules several tens of nanometers in size. An electron microscope can also obtain the images of materials with a size of several nanometers, but get structural information only. On the other hand, the super-resolution microscope uses a fluorescent signal to label multiple biomolecules and obtain multiplexed targets for understanding their variations.
Among the methods of super-resolution microscopy, first, single-molecule localization microscopy (SMLM) is used with the blinking phenomena of fluorescent materials such as PALM, STORM, and DNA-PAINT. SMLM is used with these blinking events of fluorescent proteins or dyes. The emitters are fitted by a Gaussian function and then are found by a position of maximum intensity. Recently, researchers evaluated the performance of algorithms for super-resolution reconstruction [66]. It was also presented by the algorithms to reduce artifacts during the reconstruction process [67, 68]. The blinking events for PALM and STORM occur by activating all fluorescent proteins or dyes until they are photo bleached. And then a subset of proteins or dyes stochastically emit the fluorescent signal. The DNA-PAINT technique, however, does not need to take a bleaching step. DNA-PAINT is used by transient binding between two single-stranded DNA, called a docking strand and an imager strand.
In Fig. 2(a), the STED microscope is a method to reduce the size of an illumination beam. With a donut-shaped depletion laser around the illumination laser, the beam size can be reduced to the size of nanometers. By scanning the set of two lasers over the entire area of the sample, we can obtain a super-resolution image. Figure 2(a) shows an
In Fig. 2(b), the left side shows the total internal reflection fluorescence (TIRF) image of the lysosomal transmembrane protein, CD63, labelled with PA-FP Kaede in COS-7 cells. The right side shows the PALM image of the left side in the same region [23]. Figure 2(c) shows microtubule images by STORM in BS-C-1 cells. The left side is a conventional fluorescence image. The right side is a three-dimensional super-resolution image with the
In the DNA-PAINT technique, the imager strands are replaced by flowing in fresh imaging buffer. With this method, we solve the photobleaching problem of fluorescent dye. On the other hand, the signal-to-noise ratio (SNR) is lower than that of STORM or PALM because the unbound imager strand floats on the imaging buffer. With the principle of the super-resolution microscope, there are many blinking events to reduce the time to acquire a super-resolved image. In DNA-PAINT, however, increasing the concentration of the imager strand results in decreasing the SNR, so it is not possible to increase the concentration of the imager strand unlimitedly for the purpose of increasing the number of blinking events. Therefore, it takes a few minutes to tens of minutes to obtain a super-resolved image by DNA-PAINT.
The FRET-PAINT technique was a method to solve the disadvantages of DNA-PAINT, which takes a long time to obtain a super-resolution image. FRET is an abbreviation of fluorescence resonance energy transfer. It refers to a phenomenon in which energy is transferred when two fluorescent molecules are close by a few nanometers. In particular, the wavelength of the excitation laser is far distant to that of fluorescent signal. If bandpass filters and dichroic mirrors are used, we can precisely obtain the fluorescent signal. In addition, the method of TIRF using an evanescent wave can be applied to biophysics and biochemistry by single-molecule localization detection. In the range of several nanometers for the FRET phenomenon, it is used to study protein-protein, protein-DNA, or protein-RNA interactions.
It was presented with two principles of FRET-PAINT for super-resolution imaging. The first principle is that the docking strand is not labeled with a fluorescent dye, but the acceptor strand and the donor strand are labeled by different fluorescent dyes. If the acceptor and the donor are simultaneously bound to the docking strand, the FRET phenomenon occurs. The second principle is that a fluorescent dye is labeled on the docking strand as an acceptor, and an imager strand as a donor is bound to the FRET phenomenon. In Fig. 2(e), the acquisition time to obtain a super-resolution image was decreased to a few seconds with the first principle of FRET-PAINT.
Structured-illumination microscopy (SIM) breaks through the diffraction limit with spatial frequency using an illumination pattern. SIM can only be applied to a conventional wide-field microscope and a single laser. The resolution of SIM is reduced by a half of the resolution of the wide-field microscope. Figure 2(f) shows microtubule images in a S2 cell with a conventional TIRF microscope (upper panel) and TIRF-SIM (lower panel) [15].
As discussed in the previous section, the super-resolution microscope has mainly been used to obtain images of biomolecules in cells with a size of tens of nanometers. As these techniques advanced, researchers addressed the acquisition of super-resolution images from volumetric samples such as the embryo of
For volumetric imaging with a super-resolution microscope, localization precision, which indicates the performance of super-resolution microscopy, is important. Localization precision means how accurately the position of the emitter detected in the camera can be determined. Thompson
Figure 3(a) shows single-molecule images of the nuclear pore complexes in cells with highly inclined and laminated optical sheet (HILO) microscopy [80]. In order to obtain single-molecule images
Another method for volumetric imaging is a lattice light sheet microscope. In general, the beam from a laser is shaped by a Gaussian beam. However, it is difficult to illuminate the depth of the tissue sample with the Gaussian beam due to light scattering. In a square lattice beam, the point spread function becomes sharper than a Gaussian illumination beam. Figure 3(b) shows a chromosomal passenger protein GFP–AIR-2 (green) and plasma membranes and histones (red) in embryos of
The point spread function engineering method is applied to obtain a volumetric super-resolution image. In this method, the point spread function is transformed by a deformable mirror or a spatial light modulator (SLM) to make it more robust to scattering and aberration. Figure 3(c) shows a three-dimensional super-resolution image of microtubules with the self-bending point spread function (SB-PSF) [84]. In this study, the SB-PSF was designed by no side-lobe and detected by three-dimensional localization of emitters. By solving the diffraction and propagation-dependent bending phenomenon, precise volumetric images can be obtained.
Figure 3(d) shows super-resolution images of microtubules and mitochondria using a spinning disk confocal microscope. In this method, a disk with a pinhole array is installed in the pathway of the excitation laser and the objective lens. The image is acquired by spinning this disk. It was difficult to obtain a whole-cell image with the conventional TIRF or HILO method. Using the spinning disk confocal microscope, however, super-resolution images were obtained from 6-μm-thick HeLa cells at 500-nm intervals [85].
The performance of multi-emitter fitting is improved by introducing machine learning to PSF engineering. Figure 3(e) shows a three-dimensional super-resolution image of mitochondria over a 4 μm axial range [86]. In this method, the tetrapod-shaped PSF was engineered by the phase mask using an SLM as a function of axial position. Overlapping emitters with the reshaped PSF were analyzed by the convolutional neural network. The researchers designed an optimal PSF for three-dimensional imaging of a large axial range.
Figure 3(f) shows a three-dimensional image of mitochondria by labeling with TOM22 antibody in a COS-7 cell with the Zernike optimized localization approach in 3D (ZOLA-3D) [87]. In this technique, the researchers developed a phase retrieval method with a deformable mirror. The phase of PSF was calculated by Zernike polynomials. The position of the emitter was predicted numerically by the Zernike coefficients. Using ZOLA-3D, a three-dimensional super-resolution image was obtained in the range of 2 μm in a single imaging plane.
The line-scan confocal microscope shortened the time required to acquire a single image by scanning the beam as a line shape instead of the conventional point scan [88, 89]. With the confocal technique in this method, the SNR is increased by blocking the fluorescence signal out of the focal plane. Figure 3(g) shows a three-dimensional image for the synapse of 100-μm-thick mouse brain tissue. The presynapse is targeted by the Bassoon protein (green) and the postsynapse is targeted by the Homer1 protein (red) [90].
Figure 3(h) shows a three-dimensional super-resolution image of the nuclear lamina (Lamin B1) of a HeLa cell with a tilted light sheet microscope [91]. Compared to conventional light sheet illumination, tilted light sheet illumination has an advantage of volumetric imaging to extend the range of PSF in the axial direction. By reshaping the double-helix phase pattern on the PSF, the precise axial position of the emitter can be obtained.
In order to take a deep tissue image using a fluorescence microscope, aberration in the sample should be corrected to obtain a fluorescence image from a deep position. A novel technique for this aberration correction is adaptive optics. There are two methods for adaptive optics: (1) direct sensing and (2) indirect sensing of aberrations. The direct sensing method uses a wavefront sensing instrument such as a Shack-Hartman sensor. The aberration is directly evaluated by the wavefront sensor. In the indirect sensing method, the sample-induced aberration is calculated by image metrics such as intensity, contrast, or sharpness. This process is optimized by aberration correction using image-based metrics [92–94].
Figure 4 shows two principles of adaptive optics. First, Fig. 4(a) is a method of directly sensing sample-induced aberration using a wavefront sensor. Aberration of the scattering media is measured with a guide star inside a sample. Next, we measure the fluorescent signal near the guide star, and then correct the aberration using the information obtained by the guide star. This method, previously used in astronomy, was developed to compensate for the aberration caused by the Earth’s atmosphere when observing signals on the surface. As to imaging biological samples with a microscope, the guide stars are used by the nanoparticles, fluorescent beads, or exogenous agents. The guide stars, however, should be near the region of interest. The imaging area is conventionally determined by the isoplanatic patches adjacent to guide stars. The equipment for the wavefront sensor is also expensive. There is a technical demand to use it for practical experiments.
Figure 4(b) shows an indirect sensing method with a camera or photomultiplier tube. The sample-induced aberration is calculated by an image metric such as the intensity or sharpness of the detected image. The wavefront shaping device is controlled iteratively in the direction of maximizing the value of the image metric. This indirect method does not need to measure the signal of the guide star. Therefore, the sample preparation and equipment are simpler than that of the direct sensing method. The instruments of the image metric are integrated easily and inexpensively. The iterative calculation and update process of the image metric, however, is a time-consuming operation. This method is also less accurate than the direct sensing method.
In this section, we review several studies for the super-resolution microscope with adaptive optics for volumetric imaging. The results of these studies were obtained by adaptive optics with super-resolution microscopy such as PALM, iPALM, STORM, and STED.
Figure 5(a) shows an image of synaptonemal complexes in a mouse spermatocyte using two deformable mirrors with iPALM technology [95]. In this study, the fluorescent signals were detected by two objective lenses. Both signals of the interference path were calculated, respectively. For volumetric imaging, the PSF should be delivered to the axial position of the sample. The shape of the PSF, however, is deformed due to sample-induced aberration. Two deformable mirrors were installed in two interferometric cavities to compensate for the sample-induced aberration depending on the sample-to-sample variation and the depth of field. To obtain three-dimensional images, astigmatism was also applied with the deformable mirrors without an additional lens such as a cylindrical lens.
Another volumetric imaging method is a combination of PSF engineering and aberration correction. Figure 5(b) shows three-dimensional images of fibrillar amyloid-β plaques in a 30-μm section of a mouse tissue sample [96]. In this study, the researchers developed a technique for aberration correction in different depths. They also modulated the shape of PSF in the axial direction with a deformable mirror. In a previous study, the Zhuang group reported two-dimensional images from a brain slice using a single-molecule localization microscope and then computed the images to obtain three-dimensional images [97]. The Huang group, however, acquired volumetric super-resolution images from a single thick brain tissue with adaptive optics and PSF engineering.
A STED microscope with adaptive optics is applied to acquire volumetric images [98, 99]. In this technique, the STED microscope consists of three paths: a depletion path, excitation path, and emission path. All three paths are affected by sample-induced aberration. The STED microscope combined with adaptive optics, therefore, requires the correction of all paths. A deformable mirror is used to compensate for the sample-induced aberration in the common path of the microscope. An SLM is also applied to correct the aberration of the depletion beam. Figure 5(c) shows a super-resolution image of a vesicular glutamate transporter in
Many methods of super-resolution microscopy have been applied to fixed cells or tissues
Figure 6(b) shows 100-μm-thick widefield images of motor neurons in zebrafish
Figure 6(c) shows the adherens junctions (AJs) in the anterior bulb of
With the super-resolution microscope, we can observe biomolecules of several tens of nanometers beyond the diffraction limit. The diffraction-limited images were obtained using this microscope as targets for microstructures such as microtubules, actin, and mitochondria in cells. Studies were also conducted to obtain super-resolution images from tissue samples such as mouse brains. In tissue samples, sample-induced aberration including photon scattering was the technical barrier for super-resolution imaging. Adaptive optics was applied to the super-resolution microscope to break through this barrier.
Volumetric imaging can be performed by combining adaptive optics with super-resolution imaging technology. In previous studies, the techniques for volumetric imaging without adaptive optics were multi-angle interference [104], oblique plane excitation [105], and multi-focal excitation [106]. However, it is more promising to apply adaptive optics in that multiple aberrations including sample-induced aberration should be corrected in tissue samples. Methods of volumetric imaging with adaptive optics were studied with SMLM [95, 96, 107, 108], SIM [99, 100], and STED microscopes [11, 102, 109]. The techniques of adaptive optics continue to be developed with the sensor-based method, the sensorless method, and the post-processing method. Methods for volumetric imaging are also being developed by combining these technologies with the super-resolution microscope.
Many researchers have performed experiments in the direction of decreasing the resolution of super-resolution microscopes such as MINFLUX [110–112]. However, it is still not good compared to the resolution of an electron microscope. If two microscope technologies are combined, an image of a specific target can be obtained at nanometer scale. The images of multiple targets can be obtained simultaneously by applying multicolor imaging methods, while an image can be obtained by an electron microscope at nanometer scales.
Recently, a novel method was reported where the cellular ultrastructure of specific biomolecules was preserved by epoxy resin embedding for a correlative super-resolution microscope and an electron microscope [113]. In addition, a cryogenic super-resolution microscope was developed and the researchers obtained multicolor three-dimensional super-resolution images at 10 K temperature. They used SMLM and SIM with resolutions of 4 nm to 8 nm [114].
One of the disadvantages of super-resolution microscopy is that it takes a long time to acquire an image. For example, it took more than a few minutes to get a single super-resolved image with PALM or STORM. With the DNA-PAINT method, which can solve the photobleaching problem, it took several tens of minutes to obtain one super-resolution image. To speed up the imaging time, FRET-PAINT technology using FRET phenomena was developed. Recently, a study reported that researchers developed a high-speed super-resolution imaging method with protein-assisted DNA-PAINT, called Ago-PAINT. According to this method, the Argonaute protein was bound to the imager strand to create a DNA-protein complex before imaging. When this complex was used for super-resolution imaging, the imaging speed was 10 times faster than the conventional DNA-PAINT method [115]. In another study, periodic DNA sequence motifs were used to achieve a 100 times faster sampling rate than conventional DNA-PAINT [116].
The super-resolution microscope has emerged as a method for observing biomolecules with a size of several tens of nanometers beyond the diffraction limit. Studies on super-resolution imaging have been performed in the direction of improving the performance of this microscope. In order to obtain images of several nanometer-sized materials, the resolution is decreased. By reducing the imaging time, a super-resolution image was obtained in seconds. In addition, by combining an electron microscope and a super-resolution microscope, three-dimensional images are obtained at very low temperatures. The study also reported that the time to analyze the raw data was reduced using machine learning [86, 117–125]. Super-resolution microscopes have continuously evolved and are used to solve biological problems. The super-resolution microscope, therefore, can be applied to neuroscience including human connectome and biomedical research for disease treatment [126].
The authors declare no conflicts of interest.
No data were generated or analyzed in the presented research.
FIG. 1 and FIG. 4 were created with BioRender.com.
National Research Foundation of Korea (NRF-2019R1I1 A1A01059015, NRF-2022R1A6A3A01085960, NRF-2020R1A6A3A01100231, NRF-2019R1C1C1008175, NRF-2021R1A4A5028966)
Curr. Opt. Photon. 2022; 6(6): 550-564
Published online December 25, 2022 https://doi.org/10.3807/COPP.2022.6.6.550
Copyright © Optical Society of Korea.
Sangjun Park1,2, Cheol Hong Min1,2, Seokyoung Han3, Eunjin Choi1,2, Kyung-Ok Cho2,4, Hyun-Jong Jang2,5, Moonseok Kim1,2
1Department of Medical Life Sciences, College of Medicine, The Catholic University of Korea, Seoul 06591, Korea
2Department of Biomedicine & Health Sciences, College of Medicine, The Catholic University of Korea, Seoul 06591, Korea
3Department of Mechanical Engineering, University of Louisville, Louisville, Kentucky 40208, USA
4Department of Pharmacology, College of Medicine, The Catholic University of Korea, Seoul 06591, Korea
5Department of Physiology, College of Medicine, The Catholic University of Korea, Seoul 06591, Korea
Correspondence to:*moonseok@catholic.ac.kr, ORCID 0000-0003-3100-8980
†These authors contributed equally to this paper.
This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http://creativecommons.org/licenses/by-nc/4.0/) which permits unrestricted non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited.
Optical microscopy is a useful tool for study in the biological sciences. With an optical microscope, we can observe the micro world of life such as tissues, cells, and proteins. A fluorescent dye or a fluorescent protein provides an opportunity to mark a specific target in the crowd of biological samples, so that an image of a specific target can be observed by an optical microscope. The optical microscope, however, is constrained in resolution due to diffraction limit. Super-resolution microscopy made a breakthrough with this diffraction limit. Using a super-resolution microscope, many biomolecules are observed beyond the diffraction limit in cells. In the case of volumetric imaging, the super-resolution techniques are only applied to a limited area due to long imaging time, multiple scattering of photons, and sample-induced aberration in deep tissue. In this article, we review recent advances in superresolution microscopy for volumetric imaging. The super-resolution techniques have been integrated with various modalities, such as a line-scan confocal microscope, a spinning disk confocal microscope, a light sheet microscope, and point spread function engineering. Super-resolution microscopy combined with adaptive optics by compensating for wave distortions is a promising method for deep tissue imaging and biomedical applications.
Keywords: Adaptive optics, Super-resolution microscope, Volumetric imaging, Wavefront shaping
Optical microscopy has long been useful to study biological sciences. With an optical microscope, we take an image and get information on the micro world. Among many proteins, DNA, and RNA in a biological sample, fluorescent dye or a fluorescent protein provides an opportunity to mark a specific target. The optical resolution, however, is limited by diffraction. The diffraction limit is known as the Abbe limit as in the following equation [1]:
where
As shown in Fig. 1, mammalian cells and
The techniques of the super-resolution microscope are presented in several categories theoretically and experimentally. The major techniques, for example, are stimulated emission depletion (STED) microscope [5–12], structured illumination microscope (SIM) [13–22], photo-activated localization microscope (PALM) [23–28], stochastic optical reconstruction microscope (STORM) [29–37], fluorescence photo-activated localization microscope (FPALM) [38–40], interferometric photoactivated localization microscopy (iPALM) [41, 42], 4Pi single marker switching nanoscopy (4Pi-SMSN) [43], reversible saturable optical linear fluorescence transitions microscope (RESOLFT) [44–48], super-resolution optical fluctuation imaging (SOFI) [49–54], DNA point accumulation for imaging in nanoscale topography (DNA-PAINT) [55–61], and FRET-based DNA-PAINT (FRET-PAINT) [62–65].
In this article, we review the methods of super-resolution microscopy combined with adaptive optics for volumetric imaging. In volumetric imaging, the super-resolution microscopes have been applied to a limited area due to long imaging time, multiple scattering, and sample-induced aberration in deep tissue. To break through the technical barriers, the super-resolution microscopes have been integrated with line-scan confocal microscopy, spinning disk confocal microscopy, light sheet microscopy, or point spread function engineering. The super-resolution microscopy incorporated with adaptive optics has been exploited for volumetric imaging by compensating for the wave distortions in biological tissue. As a result of the improvement, super-resolution microscopy with adaptive optical techniques can be advantageous for deep tissue imaging and biomedical applications. Among several methods of super-resolution imaging, we extensively discuss STED, SIM, and single-molecule localization microscopy with adaptive optics in this review.
The super-resolution microscope offers the advantage of breaking the diffraction limit and obtaining images of biomolecules several tens of nanometers in size. An electron microscope can also obtain the images of materials with a size of several nanometers, but get structural information only. On the other hand, the super-resolution microscope uses a fluorescent signal to label multiple biomolecules and obtain multiplexed targets for understanding their variations.
Among the methods of super-resolution microscopy, first, single-molecule localization microscopy (SMLM) is used with the blinking phenomena of fluorescent materials such as PALM, STORM, and DNA-PAINT. SMLM is used with these blinking events of fluorescent proteins or dyes. The emitters are fitted by a Gaussian function and then are found by a position of maximum intensity. Recently, researchers evaluated the performance of algorithms for super-resolution reconstruction [66]. It was also presented by the algorithms to reduce artifacts during the reconstruction process [67, 68]. The blinking events for PALM and STORM occur by activating all fluorescent proteins or dyes until they are photo bleached. And then a subset of proteins or dyes stochastically emit the fluorescent signal. The DNA-PAINT technique, however, does not need to take a bleaching step. DNA-PAINT is used by transient binding between two single-stranded DNA, called a docking strand and an imager strand.
In Fig. 2(a), the STED microscope is a method to reduce the size of an illumination beam. With a donut-shaped depletion laser around the illumination laser, the beam size can be reduced to the size of nanometers. By scanning the set of two lasers over the entire area of the sample, we can obtain a super-resolution image. Figure 2(a) shows an
In Fig. 2(b), the left side shows the total internal reflection fluorescence (TIRF) image of the lysosomal transmembrane protein, CD63, labelled with PA-FP Kaede in COS-7 cells. The right side shows the PALM image of the left side in the same region [23]. Figure 2(c) shows microtubule images by STORM in BS-C-1 cells. The left side is a conventional fluorescence image. The right side is a three-dimensional super-resolution image with the
In the DNA-PAINT technique, the imager strands are replaced by flowing in fresh imaging buffer. With this method, we solve the photobleaching problem of fluorescent dye. On the other hand, the signal-to-noise ratio (SNR) is lower than that of STORM or PALM because the unbound imager strand floats on the imaging buffer. With the principle of the super-resolution microscope, there are many blinking events to reduce the time to acquire a super-resolved image. In DNA-PAINT, however, increasing the concentration of the imager strand results in decreasing the SNR, so it is not possible to increase the concentration of the imager strand unlimitedly for the purpose of increasing the number of blinking events. Therefore, it takes a few minutes to tens of minutes to obtain a super-resolved image by DNA-PAINT.
The FRET-PAINT technique was a method to solve the disadvantages of DNA-PAINT, which takes a long time to obtain a super-resolution image. FRET is an abbreviation of fluorescence resonance energy transfer. It refers to a phenomenon in which energy is transferred when two fluorescent molecules are close by a few nanometers. In particular, the wavelength of the excitation laser is far distant to that of fluorescent signal. If bandpass filters and dichroic mirrors are used, we can precisely obtain the fluorescent signal. In addition, the method of TIRF using an evanescent wave can be applied to biophysics and biochemistry by single-molecule localization detection. In the range of several nanometers for the FRET phenomenon, it is used to study protein-protein, protein-DNA, or protein-RNA interactions.
It was presented with two principles of FRET-PAINT for super-resolution imaging. The first principle is that the docking strand is not labeled with a fluorescent dye, but the acceptor strand and the donor strand are labeled by different fluorescent dyes. If the acceptor and the donor are simultaneously bound to the docking strand, the FRET phenomenon occurs. The second principle is that a fluorescent dye is labeled on the docking strand as an acceptor, and an imager strand as a donor is bound to the FRET phenomenon. In Fig. 2(e), the acquisition time to obtain a super-resolution image was decreased to a few seconds with the first principle of FRET-PAINT.
Structured-illumination microscopy (SIM) breaks through the diffraction limit with spatial frequency using an illumination pattern. SIM can only be applied to a conventional wide-field microscope and a single laser. The resolution of SIM is reduced by a half of the resolution of the wide-field microscope. Figure 2(f) shows microtubule images in a S2 cell with a conventional TIRF microscope (upper panel) and TIRF-SIM (lower panel) [15].
As discussed in the previous section, the super-resolution microscope has mainly been used to obtain images of biomolecules in cells with a size of tens of nanometers. As these techniques advanced, researchers addressed the acquisition of super-resolution images from volumetric samples such as the embryo of
For volumetric imaging with a super-resolution microscope, localization precision, which indicates the performance of super-resolution microscopy, is important. Localization precision means how accurately the position of the emitter detected in the camera can be determined. Thompson
Figure 3(a) shows single-molecule images of the nuclear pore complexes in cells with highly inclined and laminated optical sheet (HILO) microscopy [80]. In order to obtain single-molecule images
Another method for volumetric imaging is a lattice light sheet microscope. In general, the beam from a laser is shaped by a Gaussian beam. However, it is difficult to illuminate the depth of the tissue sample with the Gaussian beam due to light scattering. In a square lattice beam, the point spread function becomes sharper than a Gaussian illumination beam. Figure 3(b) shows a chromosomal passenger protein GFP–AIR-2 (green) and plasma membranes and histones (red) in embryos of
The point spread function engineering method is applied to obtain a volumetric super-resolution image. In this method, the point spread function is transformed by a deformable mirror or a spatial light modulator (SLM) to make it more robust to scattering and aberration. Figure 3(c) shows a three-dimensional super-resolution image of microtubules with the self-bending point spread function (SB-PSF) [84]. In this study, the SB-PSF was designed by no side-lobe and detected by three-dimensional localization of emitters. By solving the diffraction and propagation-dependent bending phenomenon, precise volumetric images can be obtained.
Figure 3(d) shows super-resolution images of microtubules and mitochondria using a spinning disk confocal microscope. In this method, a disk with a pinhole array is installed in the pathway of the excitation laser and the objective lens. The image is acquired by spinning this disk. It was difficult to obtain a whole-cell image with the conventional TIRF or HILO method. Using the spinning disk confocal microscope, however, super-resolution images were obtained from 6-μm-thick HeLa cells at 500-nm intervals [85].
The performance of multi-emitter fitting is improved by introducing machine learning to PSF engineering. Figure 3(e) shows a three-dimensional super-resolution image of mitochondria over a 4 μm axial range [86]. In this method, the tetrapod-shaped PSF was engineered by the phase mask using an SLM as a function of axial position. Overlapping emitters with the reshaped PSF were analyzed by the convolutional neural network. The researchers designed an optimal PSF for three-dimensional imaging of a large axial range.
Figure 3(f) shows a three-dimensional image of mitochondria by labeling with TOM22 antibody in a COS-7 cell with the Zernike optimized localization approach in 3D (ZOLA-3D) [87]. In this technique, the researchers developed a phase retrieval method with a deformable mirror. The phase of PSF was calculated by Zernike polynomials. The position of the emitter was predicted numerically by the Zernike coefficients. Using ZOLA-3D, a three-dimensional super-resolution image was obtained in the range of 2 μm in a single imaging plane.
The line-scan confocal microscope shortened the time required to acquire a single image by scanning the beam as a line shape instead of the conventional point scan [88, 89]. With the confocal technique in this method, the SNR is increased by blocking the fluorescence signal out of the focal plane. Figure 3(g) shows a three-dimensional image for the synapse of 100-μm-thick mouse brain tissue. The presynapse is targeted by the Bassoon protein (green) and the postsynapse is targeted by the Homer1 protein (red) [90].
Figure 3(h) shows a three-dimensional super-resolution image of the nuclear lamina (Lamin B1) of a HeLa cell with a tilted light sheet microscope [91]. Compared to conventional light sheet illumination, tilted light sheet illumination has an advantage of volumetric imaging to extend the range of PSF in the axial direction. By reshaping the double-helix phase pattern on the PSF, the precise axial position of the emitter can be obtained.
In order to take a deep tissue image using a fluorescence microscope, aberration in the sample should be corrected to obtain a fluorescence image from a deep position. A novel technique for this aberration correction is adaptive optics. There are two methods for adaptive optics: (1) direct sensing and (2) indirect sensing of aberrations. The direct sensing method uses a wavefront sensing instrument such as a Shack-Hartman sensor. The aberration is directly evaluated by the wavefront sensor. In the indirect sensing method, the sample-induced aberration is calculated by image metrics such as intensity, contrast, or sharpness. This process is optimized by aberration correction using image-based metrics [92–94].
Figure 4 shows two principles of adaptive optics. First, Fig. 4(a) is a method of directly sensing sample-induced aberration using a wavefront sensor. Aberration of the scattering media is measured with a guide star inside a sample. Next, we measure the fluorescent signal near the guide star, and then correct the aberration using the information obtained by the guide star. This method, previously used in astronomy, was developed to compensate for the aberration caused by the Earth’s atmosphere when observing signals on the surface. As to imaging biological samples with a microscope, the guide stars are used by the nanoparticles, fluorescent beads, or exogenous agents. The guide stars, however, should be near the region of interest. The imaging area is conventionally determined by the isoplanatic patches adjacent to guide stars. The equipment for the wavefront sensor is also expensive. There is a technical demand to use it for practical experiments.
Figure 4(b) shows an indirect sensing method with a camera or photomultiplier tube. The sample-induced aberration is calculated by an image metric such as the intensity or sharpness of the detected image. The wavefront shaping device is controlled iteratively in the direction of maximizing the value of the image metric. This indirect method does not need to measure the signal of the guide star. Therefore, the sample preparation and equipment are simpler than that of the direct sensing method. The instruments of the image metric are integrated easily and inexpensively. The iterative calculation and update process of the image metric, however, is a time-consuming operation. This method is also less accurate than the direct sensing method.
In this section, we review several studies for the super-resolution microscope with adaptive optics for volumetric imaging. The results of these studies were obtained by adaptive optics with super-resolution microscopy such as PALM, iPALM, STORM, and STED.
Figure 5(a) shows an image of synaptonemal complexes in a mouse spermatocyte using two deformable mirrors with iPALM technology [95]. In this study, the fluorescent signals were detected by two objective lenses. Both signals of the interference path were calculated, respectively. For volumetric imaging, the PSF should be delivered to the axial position of the sample. The shape of the PSF, however, is deformed due to sample-induced aberration. Two deformable mirrors were installed in two interferometric cavities to compensate for the sample-induced aberration depending on the sample-to-sample variation and the depth of field. To obtain three-dimensional images, astigmatism was also applied with the deformable mirrors without an additional lens such as a cylindrical lens.
Another volumetric imaging method is a combination of PSF engineering and aberration correction. Figure 5(b) shows three-dimensional images of fibrillar amyloid-β plaques in a 30-μm section of a mouse tissue sample [96]. In this study, the researchers developed a technique for aberration correction in different depths. They also modulated the shape of PSF in the axial direction with a deformable mirror. In a previous study, the Zhuang group reported two-dimensional images from a brain slice using a single-molecule localization microscope and then computed the images to obtain three-dimensional images [97]. The Huang group, however, acquired volumetric super-resolution images from a single thick brain tissue with adaptive optics and PSF engineering.
A STED microscope with adaptive optics is applied to acquire volumetric images [98, 99]. In this technique, the STED microscope consists of three paths: a depletion path, excitation path, and emission path. All three paths are affected by sample-induced aberration. The STED microscope combined with adaptive optics, therefore, requires the correction of all paths. A deformable mirror is used to compensate for the sample-induced aberration in the common path of the microscope. An SLM is also applied to correct the aberration of the depletion beam. Figure 5(c) shows a super-resolution image of a vesicular glutamate transporter in
Many methods of super-resolution microscopy have been applied to fixed cells or tissues
Figure 6(b) shows 100-μm-thick widefield images of motor neurons in zebrafish
Figure 6(c) shows the adherens junctions (AJs) in the anterior bulb of
With the super-resolution microscope, we can observe biomolecules of several tens of nanometers beyond the diffraction limit. The diffraction-limited images were obtained using this microscope as targets for microstructures such as microtubules, actin, and mitochondria in cells. Studies were also conducted to obtain super-resolution images from tissue samples such as mouse brains. In tissue samples, sample-induced aberration including photon scattering was the technical barrier for super-resolution imaging. Adaptive optics was applied to the super-resolution microscope to break through this barrier.
Volumetric imaging can be performed by combining adaptive optics with super-resolution imaging technology. In previous studies, the techniques for volumetric imaging without adaptive optics were multi-angle interference [104], oblique plane excitation [105], and multi-focal excitation [106]. However, it is more promising to apply adaptive optics in that multiple aberrations including sample-induced aberration should be corrected in tissue samples. Methods of volumetric imaging with adaptive optics were studied with SMLM [95, 96, 107, 108], SIM [99, 100], and STED microscopes [11, 102, 109]. The techniques of adaptive optics continue to be developed with the sensor-based method, the sensorless method, and the post-processing method. Methods for volumetric imaging are also being developed by combining these technologies with the super-resolution microscope.
Many researchers have performed experiments in the direction of decreasing the resolution of super-resolution microscopes such as MINFLUX [110–112]. However, it is still not good compared to the resolution of an electron microscope. If two microscope technologies are combined, an image of a specific target can be obtained at nanometer scale. The images of multiple targets can be obtained simultaneously by applying multicolor imaging methods, while an image can be obtained by an electron microscope at nanometer scales.
Recently, a novel method was reported where the cellular ultrastructure of specific biomolecules was preserved by epoxy resin embedding for a correlative super-resolution microscope and an electron microscope [113]. In addition, a cryogenic super-resolution microscope was developed and the researchers obtained multicolor three-dimensional super-resolution images at 10 K temperature. They used SMLM and SIM with resolutions of 4 nm to 8 nm [114].
One of the disadvantages of super-resolution microscopy is that it takes a long time to acquire an image. For example, it took more than a few minutes to get a single super-resolved image with PALM or STORM. With the DNA-PAINT method, which can solve the photobleaching problem, it took several tens of minutes to obtain one super-resolution image. To speed up the imaging time, FRET-PAINT technology using FRET phenomena was developed. Recently, a study reported that researchers developed a high-speed super-resolution imaging method with protein-assisted DNA-PAINT, called Ago-PAINT. According to this method, the Argonaute protein was bound to the imager strand to create a DNA-protein complex before imaging. When this complex was used for super-resolution imaging, the imaging speed was 10 times faster than the conventional DNA-PAINT method [115]. In another study, periodic DNA sequence motifs were used to achieve a 100 times faster sampling rate than conventional DNA-PAINT [116].
The super-resolution microscope has emerged as a method for observing biomolecules with a size of several tens of nanometers beyond the diffraction limit. Studies on super-resolution imaging have been performed in the direction of improving the performance of this microscope. In order to obtain images of several nanometer-sized materials, the resolution is decreased. By reducing the imaging time, a super-resolution image was obtained in seconds. In addition, by combining an electron microscope and a super-resolution microscope, three-dimensional images are obtained at very low temperatures. The study also reported that the time to analyze the raw data was reduced using machine learning [86, 117–125]. Super-resolution microscopes have continuously evolved and are used to solve biological problems. The super-resolution microscope, therefore, can be applied to neuroscience including human connectome and biomedical research for disease treatment [126].
The authors declare no conflicts of interest.
No data were generated or analyzed in the presented research.
FIG. 1 and FIG. 4 were created with BioRender.com.
National Research Foundation of Korea (NRF-2019R1I1 A1A01059015, NRF-2022R1A6A3A01085960, NRF-2020R1A6A3A01100231, NRF-2019R1C1C1008175, NRF-2021R1A4A5028966)