Periodic Reporting for period 1 - SM-SPAD (Single-molecule imaging with SPAD array detection)
Reporting period: 2021-03-01 to 2023-02-28
Microscopes have become a fundamental tool ubiquitous with high-quality biological research, and have enabled the visualization of cells, cell organelles, and even single molecules. However, the resolving power of conventional widefield microscopes is limited to about 250 nm due to the diffraction of light. Consequently, two structures closer than this limit are unresolvable from one another. This limit is explicitly detrimental to structural biology studies, where the species of interest, typically proteins, nucleic acids, or other macromolecules, is much smaller than 100 nm. Our understanding of the organization of these structures and the way they generate responses in cells is therefore currently restricted.
Several imaging techniques that are able to circumvent the diffraction limit have recently been introduced in fluorescence microscopy. Single-molecule (SM) localization microscopy was demonstrated to yield a lateral resolution of 20 nm by sequentially and stochastically switching on single fluorescent molecules and determining their position in the image plane by analyzing the observed fluorescence. However, these techniques require a lot of emission photons to get a reasonable localization precision. Unfortunately, the fluorescence emission is limited by photobleaching of the fluorophores.
To overcome this problem, the concept of MINFLUX was recently introduced. MINFLUX is a technique in which a doughnut-shaped spot of light is used to find the position of a single fluorescence-tagged molecule. By moving the spot so that the center of the doughnut is in close proximity to the position of the molecule, only a minimum dose of fluorescence is generated. A triangulation procedure that is realized by slightly moving the excitation beam allows estimation of the molecule’s location with very high accuracy and minimum photon flux. Although MINFLUX provides a significantly better resolution compared to standard localization microscopy, it still neglects crucial information needed to improve the resolution further and speed up the imaging process, namely: (i) the point-detection system does not allow directly recording the image of the fluorescent molecule in the image plane, thus a rather complex beam-scanning approach is required; (ii) it does not leverage the fluorescence lifetime information, which can provide nanometer-scale information on the structure of interest. Recording this information is crucial for understanding the organization of molecular complexes. In this project, both of these limitations were tackled.
The main goal of the SM-SPAD project is to develop a 3D single-molecule fluorescence lifetime imaging technique for structural biology. The goal will be reached by combining the concept of MINFLUX with three new ideas: (i) 3D motionless structured illumination and structured detection for improved spatial resolution; (ii) fluorescence antibunching analysis to speed up the data acquisition by enabling simultaneous localization of several active molecules; (iii) SM level fluorescence lifetime analysis to extract the maximum amount of information from the sample.
The proposed molecular-scale imaging technique is perfectly suited for the study of complex biological samples. SM-SPAD is particularly promising in the field of neuroscience research in which the organization of large protein complexes are of high interest, as they may play a crucial role in the development of neurodegenerative diseases, such as Alzheimer’s and Parkinson’s.
Several imaging techniques that are able to circumvent the diffraction limit have recently been introduced in fluorescence microscopy. Single-molecule (SM) localization microscopy was demonstrated to yield a lateral resolution of 20 nm by sequentially and stochastically switching on single fluorescent molecules and determining their position in the image plane by analyzing the observed fluorescence. However, these techniques require a lot of emission photons to get a reasonable localization precision. Unfortunately, the fluorescence emission is limited by photobleaching of the fluorophores.
To overcome this problem, the concept of MINFLUX was recently introduced. MINFLUX is a technique in which a doughnut-shaped spot of light is used to find the position of a single fluorescence-tagged molecule. By moving the spot so that the center of the doughnut is in close proximity to the position of the molecule, only a minimum dose of fluorescence is generated. A triangulation procedure that is realized by slightly moving the excitation beam allows estimation of the molecule’s location with very high accuracy and minimum photon flux. Although MINFLUX provides a significantly better resolution compared to standard localization microscopy, it still neglects crucial information needed to improve the resolution further and speed up the imaging process, namely: (i) the point-detection system does not allow directly recording the image of the fluorescent molecule in the image plane, thus a rather complex beam-scanning approach is required; (ii) it does not leverage the fluorescence lifetime information, which can provide nanometer-scale information on the structure of interest. Recording this information is crucial for understanding the organization of molecular complexes. In this project, both of these limitations were tackled.
The main goal of the SM-SPAD project is to develop a 3D single-molecule fluorescence lifetime imaging technique for structural biology. The goal will be reached by combining the concept of MINFLUX with three new ideas: (i) 3D motionless structured illumination and structured detection for improved spatial resolution; (ii) fluorescence antibunching analysis to speed up the data acquisition by enabling simultaneous localization of several active molecules; (iii) SM level fluorescence lifetime analysis to extract the maximum amount of information from the sample.
The proposed molecular-scale imaging technique is perfectly suited for the study of complex biological samples. SM-SPAD is particularly promising in the field of neuroscience research in which the organization of large protein complexes are of high interest, as they may play a crucial role in the development of neurodegenerative diseases, such as Alzheimer’s and Parkinson’s.
A new optical microscope was built consisting of a picosecond MHz pulsed excitation laser. A beam splitter was used to split a single pulse of the excitation laser into several ‘sub’pulses in such a way that they are evenly distributed over the time interval between two ‘fundamental’ laser pulses. Besides the shift in time, the four pulses could also be focused at slightly different positions in space, thanks to a separate set of mirrors for each of the four beams. The setup was used with PAINT-based blinking samples. A commercial SPAD array detector was used to image the resulting fluorescence emission patterns. This detector consists of an array of 7x7 pixels, of which the central 5x5 pixels were used for the analysis. Based on the combination of laser beam position and fluorescence intensity in each pixel, the position of a fluorophore could be calculated with sub-diffraction precision. During the course of the project, a novel variant of this idea was published by Masullo et al., in which the laser beam was not split into four spatially and temporally separated beams, but instead a single beam was raster-scanned as in a confocal laser-scanning microscope. Given the high potential of this technique, it was decided to also include this modality in the SM-SPAD project. A set of galvo scan mirrors was installed to move the laser beam in a raster pattern. This also provided a convenient way of enlarging the field-of-view of the system.
Fluorescence antibunching analysis allows for speeding up the data acquisition by alleviating the sparsity constraint. To this end, a time-tagging module was needed that can measure and store the arrival times of individual photons detected by the SPAD array detector. An FPGA-based module called TTM (time-tagging module) was developed that – in a first phase – allowed the simultaneous and asynchronous photon time-tagging of 25 channels simultaneously, thus ideally suited to be combined with a 5x5 SPAD array detector. In a later phase, this platform was updated to be able to measure photon arrival times for up to 49 channels simultaneously.
However, simulations were done that showed that antibunching analysis is not the only way of understanding how many emitters are simultaneously active. Instead, by comparing the experimental micro-images made with the SPAD detector with the theoretical one for a single emitter, two emitters, or three emitters, one can with good precision predict the correct number. Once this is known, one can use a maximum likelihood estimation to find the actual position of all emitters simultaneously.
Fluorescence antibunching analysis allows for speeding up the data acquisition by alleviating the sparsity constraint. To this end, a time-tagging module was needed that can measure and store the arrival times of individual photons detected by the SPAD array detector. An FPGA-based module called TTM (time-tagging module) was developed that – in a first phase – allowed the simultaneous and asynchronous photon time-tagging of 25 channels simultaneously, thus ideally suited to be combined with a 5x5 SPAD array detector. In a later phase, this platform was updated to be able to measure photon arrival times for up to 49 channels simultaneously.
However, simulations were done that showed that antibunching analysis is not the only way of understanding how many emitters are simultaneously active. Instead, by comparing the experimental micro-images made with the SPAD detector with the theoretical one for a single emitter, two emitters, or three emitters, one can with good precision predict the correct number. Once this is known, one can use a maximum likelihood estimation to find the actual position of all emitters simultaneously.
The outcome of the SM-SPAD project may have an important impact on the microscopy community. SPAD array detectors are now commercially available and demonstrating that these detectors can be used not only for conventional image scanning microscopy or fluorescence fluctuation spectroscopy but also for single-molecule imaging increases the commercial potential of this detector type.