Fast DNA-PAINT imaging using a deep neural network

DNA points accumulation for imaging in nanoscale topography (DNA-PAINT) is a super-resolution technique with relatively easy-to-implement multi-target imaging. However, image acquisition is slow as sufficient statistical data has to be generated from spatio-temporally isolated single emitters. Here, we trained the neural network (NN) DeepSTORM to predict fluorophore positions from high emitter density DNA-PAINT data. This achieves image acquisition in one minute. We demonstrate multi-color super-resolution imaging of structure-conserved semi-thin neuronal tissue and imaging of large samples. This improvement can be integrated into any single-molecule microscope and enables fast single-molecule super-resolution microscopy.

Calibration free counting of low molecular copy numbers in single DNA-PAINT localization clusters

Single-Molecule Localization Microscopy (SMLM) has revolutionized light microscopy by enabling optical resolutions down to a few nanometer. Yet, localization precisions commonly not suffice to visually resolve single subunits in molecular assemblies or multimeric complexes. Since each targeted molecule contributes localizations during image acquisition, molecular counting approaches to reveal the target copy numbers within localization clusters have been continuously proposed since the early days of SMLM, most of which rely on preliminary knowledge of the dye photo-physics or on a calibration to a reference. Previously, we developed localization-based Fluorescence Correlation Spectroscopy (lbFCS) as an absolute ensemble counting approach for the SMLM-variant DNA-Points Accumulation for Imaging in Nanoscale Topography (PAINT), for the first time circumventing the necessity for reference calibrations. Here, we present a revised framework termed lbFCS+ which allows absolute counting of copy numbers for individual localization clusters in a single DNA-PAINT image. In lbFCS+, absolute counting in individual clusters is achieved via precise measurement of the local hybridization rates of the fluorescently-labeled oligonucleotides (imagers) employed in DNA-PAINT imaging. In proof-of-principle experiments on DNA origami nanostructures, we demonstrate the ability of lbFCS+ to truthfully determine molecular copy numbers and imager association and dissociation rates in well-separated localization clusters containing up to six docking strands. We show that lbFCS+ allows to resolve heterogeneous binding dynamics enabling the distinction of stochastically generated and a priori indistinguishable DNA assemblies. Beyond advancing quantitative DNA-PAINT imaging, we believe that lbFCS+ could find promising applications ranging from bio-sensing to DNA computing.

Direct measurement of fungal contribution to silicate weathering rates in soil

Chemical weathering produces solutes that control groundwater chemistry and supply ecosystems with essential nutrients. Although microbial activity influences silicate weathering rates and associated nutrient fluxes, its relative contribution to silicate weathering in natural settings remains largely unknown. We provide the first quantitative estimates of in situ silicate weathering rates that account for microbially induced dissolution and identify microbial actors associated with weathering. Nanoscale topography measurements showed that fungi colonizing olivine [(Mg,Fe)2SiO4] samples in a Mg-deficient forest soil accounted for up to 16% of the weathering flux after 9 mo of incubation. A local increase in olivine weathering rate was measured and attributed to fungal hyphae of Verticillium sp. Altogether, this approach provides quantitative parameters of bioweathering (i.e., rates and actors) and opens new avenues to improve elemental budgets in natural settings.

Improved resolution in single-molecule localization microscopy using QD-PAINT

Single-molecule localization microscopy (SMLM) has allowed the observation of various molecular structures in cells beyond the diffraction limit using organic dyes. In principle, the SMLM resolution depends on the precision of photoswitching fluorophore localization, which is inversely correlated with the square root of the number of photons released from the individual fluorophores. Thus, increasing the photon number by using highly bright fluorophores, such as quantum dots (QDs), can theoretically fundamentally overcome the current resolution limit of SMLM. However, the use of QDs in SMLM has been challenging because QDs have no photoswitching property, which is essential for SMLM, and they exhibit nonspecificity and multivalency, which complicate their use in fluorescence imaging. Here, we present a method to utilize QDs in SMLM to surpass the resolution limit of the current SMLM utilizing organic dyes. We confer monovalency, specificity, and photoswitchability on QDs by steric exclusion via passivation and ligand exchange with ptDNA, PEG, and casein as well as by DNA point accumulation for imaging in nanoscale topography (DNA-PAINT) via automatic thermally driven hybridization between target-bound docking and dye-bound complementary imager strands. QDs are made monovalent and photoswitchable to enable SMLM and show substantially better photophysical properties than Cy3, with higher fluorescence intensity and an improved resolution factor. QD-PAINT displays improved spatial resolution with a narrower full width at half maximum (FWHM) than DNA-PAINT with Cy3. In summary, QD-PAINT shows great promise as a next-generation SMLM method for overcoming the limited resolution of the current SMLM.