Graphical Abstract: Ribonucleoprotein (RNP) condensations through liquid-liquid phase separation play vital roles in the dynamic formation-dissolution of stress granules (SGs). These condensations are, however, usually assumed to be linked to pathologic fibrillation. Here, we show that physiologic condensation and pathologic fibrillation of RNPs are independent processes that can be unlinked with the chemical chaperone trimethylamine N-oxide (TMAO). Using the low complexity disordered domain of the archetypical SG-protein TDP-43 as model system, we show that TMAO enhances RNP liquid condensation yet inhibits protein fibrillation. Our results demonstrate effective decoupling of physiologic condensation from pathologic aggregation and suggests that selective targeting of protein fibrillation (without altering condensation) can be employed as therapeutic strategy for RNP aggregation-associated degenerative disorders.

Human NANOG expression resets stem cells to ground-state pluripotency. Here we identify the unique features of human NANOG that relate to its dose-sensitive function as a master transcription factor. NANOG is largely disordered, with a C-terminal prion-like domain that phase-transitions to gel-like condensates. Full-length NANOG readily forms higher-order oligomers at low nanomolar concentrations, orders of magnitude lower than typical amyloids. Using single-molecule Förster resonance energy transfer and fluorescence cross-correlation techniques, we show that NANOG oligomerization is essential for bridging DNA elements in vitro. Using chromatin immunoprecipitation sequencing and Hi-C 3.0 in cells, we validate that NANOG prion-like domain assembly is essential for specific DNA recognition and distant chromatin interactions. Our results provide a physical basis for the indispensable role of NANOG in shaping the pluripotent genome. NANOG’s unique ability to form prion-like assemblies could provide a cooperative and concerted DNA bridging mechanism that is essential for chromatin reorganization and dose-sensitive activation of ground-state pluripotency.

Fluorescence lifetime imaging microscopy (FLIM) is a powerful tool to quantify molecular compositions and study molecular states in complex cellular environment as the lifetime readings are not biased by fluorophore concentration or excitation power. However, the current methods to generate FLIM images are either computationally intensive or unreliable when the number of photons acquired at each pixel is low. Here we introduce a new deep learning-based method termed flimGANE (fluorescence lifetime imaging based on Generative Adversarial Network Estimation) that can rapidly generate accurate and high-quality FLIM images even in the photon-starved conditions. We demonstrated our model is up to 2,800 times faster than the gold standard time-domain maximum likelihood estimation (TD_MLE) and that flimGANE provides a more accurate analysis of low-photon-count histograms in barcode identification, cellular structure visualization, Förster resonance energy transfer characterization, and metabolic state analysis in live cells. With its advantages in speed and reliability, flimGANE is particularly useful in fundamental biological research and clinical applications, where high-speed analysis is critical.

Abstract:We first report on the development of new microscope means that reduce background contributions in fluorescence fluctuation methods: i) excitation shutter, ii) electronic switches, and iii) early and late time-gating. The elements allow for measuring molecules at low analyte concentrations. We first found conditions of early and late time-gating with time-correlated single-photon counting that made the fluorescence signal as bright as possible compared with the fluctuations in the background count rate in a diffraction-limited optical set-up. We measured about a 140-fold increase in the amplitude of autocorrelated fluorescence fluctuations at the lowest analyte concentration of about 15 pM, which gave a signal-to-background advantage of more than two-orders of magnitude. The results of this original article pave the way for single-molecule detection in solution and in live cells without immobilization or hydrodynamic/electrokinetic focusing at longer observation times than are currently available.

No abstract

Currently, work with subnanomolar concentrations is routine while femtomolar and even single-molecule studies are possible with some efforts getting high on single-molecule biophysics and biochemistry. Methodological breakthroughs, such as reducing the background light contribution in single-molecule studies, which has plagued many studies of molecular fluorescence in dilute solution, and particularly in live cells, have recently described by us. We first demonstrated how optimized time-gating of the fluorescence signal, together with time-correlated, single-photon counting, can be used to substantially boost the experimental signal-to-noise ratio about 140-fold, making it possible to measure analyte concentrations that are as low as 15 pM. By detection of femtomolar bulk concentrations, confocal microsopy has the potential to address the observation of one and the same molecule in dilute solution without immobilization or hydrodynamic/electrokinetic focusing at longer observation times than currently available. We present relevant physics. The equations are derived using Einstein's approach showing how it fits with Fick's law and the autocorrelation function. An improved technology is being developed at ISS for femtomolar microscopy. The general concepts and provided experimental examples should help to compare our approach to those used in conventional confocal microscopy.

Fluorescence lifetime imaging microscopy (FLIM) is a powerful tool to quantify molecular compositions and study the molecular states in the complex cellular environment as the lifetime readings are not biased by the fluorophore concentration or the excitation power. However, the current methods to generate FLIM images are either computationally intensive or unreliable when the number of photons acquired at each pixel is low. Here we introduce a new deep learning-based method termed flimGANE (fluorescence lifetime imaging based on Generative Adversarial Network Estimation) that can rapidly generate accurate and high-quality FLIM images even in photon-starved conditions. We demonstrated our model is not only 258 times faster than the most popular time-domain least-square estimation (TD_LSE) method but also provide more accurate analysis in barcode identification, cellular structure visualization, Förster resonance energy transfer characterization, and metabolic state analysis. With its advantages in speed and reliability, flimGANE is particularly useful in fundamental biological research and clinical applications, where ultrafast analysis is critical.

In this work, a deep learning-based method, STED-flimGANE, is introduced to achieve enhanced STED imaging resolution under a low STED-beam power and photon-starved conditions.