High-fidelity 3D live-cell nanoscopy through data-driven enhanced super-resolution radial fluctuation

Abstract

In recent years, the development of analytical approaches to super-resolution microscopy has highlighted the possibility of recovering super-resolution information from short sequences of wide-field images. Our recently developed method, SRRF (Super-Resolution Radial Fluctuations), enables long-term live-cell imaging beyond the resolution limit without specialized hardware. Here, we present eSRRF (enhanced-SRRF), a significant improvement over our initial method, enhancing image fidelity to the underlying structure and resolution. Especially, eSRRF uses automated data-driven parameter optimization, including an estimation of the number of frames necessary for optimal reconstruction. We demonstrate the improved fidelity of the images reconstructed with eSRRF, and highlight its versatility and ease of use over a wide range of microscopy techniques and biological systems. We also extend eSRRF to 3D super-resolution microscopy by combining it with multi-focus microscopy (MFM), obtaining volumetric super-resolution imaging of live cells with acquisition speed of ~1 volume/second.

Visualizing proteins by expansion microscopy

“…A third class of optical super-resolution approaches is based on determining the higher-order statistical analysis of temporal fluctuations measured in a movie, e.g. super-resolution optical fluctuation imaging (SOFI 26 ) or super-resolution radial fluctuations (SRRF 6, 7 ). The resolution of these approaches is inversely correlated to the distance between the fluorophores 6, 7, 27 and they do not require especially bright samples or special buffers, implying that they should benefit from ExM.…”

“…super-resolution optical fluctuation imaging (SOFI 26 ) or super-resolution radial fluctuations (SRRF 6, 7 ). The resolution of these approaches is inversely correlated to the distance between the fluorophores 6, 7, 27 and they do not require especially bright samples or special buffers, implying that they should benefit from ExM. To test this hypothesis, we combined X10 expansion microscopy 28, 29 with SRRF 6, 7, 27 and established a technique we term o ne-step n anoscale e xpansion (ONE) microscopy ( Extended Data Fig.…”

“…Here we report a solution to this problem, in the form of o ne-step n anoscale e xpansion (ONE) microscopy. We combined the 10-fold axial expansion of the specimen (1000-fold by volume) with a fluorescence fluctuation analysis 6, 7 to enable the description of cultured cells, tissues, viral particles, molecular complexes and single proteins. At the cellular level, using immunostaining, our technology revealed detailed nanoscale arrangements of synaptic proteins, including a quasi-regular organisation of PSD95 clusters.…”

Fluorescence fluctuation‐based super‐resolution microscopy: Basic concepts for an easy start

“…Currently, SRRF and eSRRF are available as ImageJ /FIJI plug-in NanoJ SRRF and NanoJ eSRRF, respectively (Gustafsson et al, 2016;Laine et al, 2022). In addition, SRRF has been deployed in python (Han et al, 2019), with a decrease of up to 78-folds the processing time (compared with SRRF) by allowing parallel computing supported by Compute Unified Device Architecture (CUDA) (code not available).…”

“…Currently, SRRF and eSRRF are available as ImageJ /FIJI plug‐in NanoJ SRRF and NanoJ eSRRF. 24 , 63 In addition, SRRF has been deployed in python, 59 with a decrease of up to 78‐fold for the processing time (compared with the imageJ SRRF plug‐in) by allowing parallel computing supported by compute unified device architecture (CUDA; code not available). SRRF can be performed in real‐time with parallel GPU (Graphics Processing Unit) computing with SRRF‐Stream and SRRF‐Stream+, which are only available for microscopes with specific Andor and Sona cameras.…”