Universal adaptive optics for microscopy through embedded neural network control.

The resolution and contrast of microscope imaging is often affected by aberrations introduced by imperfect optical systems and inhomogeneous refractive structures in specimens. Adaptive optics (AO) compensates these aberrations and restores diffraction limited performance. A wide range of AO solutions have been introduced, often tailored to a specific microscope type or application. Until now, a universal AO solution - one that can be readily transferred between microscope modalities - has not been deployed. We propose versatile and fast aberration correction using a physics-based machine learning assisted wavefront-sensorless AO control (MLAO) method. Unlike previous ML methods, we used a specially constructed neural network (NN) architecture, designed using physical understanding of the general microscope image formation, that was embedded in the control loop of different microscope systems. The approach means that not only is the resulting NN orders of magnitude simpler than previous NN methods, but the concept is translatable across microscope modalities. We demonstrated the method on a two-photon, a three-photon and a widefield three-dimensional (3D) structured illumination microscope. Results showed that the method outperformed commonly-used modal-based sensorless AO methods. We also showed that our ML-based method was robust in a range of challenging imaging conditions, such as 3D sample structures, specimen motion, low signal to noise ratio and activity-induced fluorescence fluctuations. Moreover, as the bespoke architecture encapsulated physical understanding of the imaging process, the internal NN configuration was no-longer a "black box", but provided physical insights on internal workings, which could influence future designs.

Systematic analysis of YFP traps reveals common mRNA/protein discordance in neural tissues.

While post-transcriptional control is thought to be required at the periphery of neurons and glia, its extent is unclear. Here, we investigate systematically the spatial distribution and expression of mRNA at single molecule sensitivity and their corresponding proteins of 200 YFP trap lines across the intact Drosophila nervous system. 97.5% of the genes studied showed discordance between the distribution of mRNA and the proteins they encode in at least one region of the nervous system. These data suggest that post-transcriptional regulation is very common, helping to explain the complexity of the nervous system. We also discovered that 68.5% of these genes have transcripts present at the periphery of neurons, with 9.5% at the glial periphery. Peripheral transcripts include many potential new regulators of neurons, glia, and their interactions. Our approach is applicable to most genes and tissues and includes powerful novel data annotation and visualization tools for post-transcriptional regulation.

Absolute quantitation of individual SARS-CoV-2 RNA molecules provides a new paradigm for infection dynamics and variant differences.

Despite an unprecedented global research effort on SARS-CoV-2, early replication events remain poorly understood. Given the clinical importance of emergent viral variants with increased transmission, there is an urgent need to understand the early stages of viral replication and transcription. We used single-molecule fluorescence in situ hybridisation (smFISH) to quantify positive sense RNA genomes with 95% detection efficiency, while simultaneously visualising negative sense genomes, subgenomic RNAs, and viral proteins. Our absolute quantification of viral RNAs and replication factories revealed that SARS-CoV-2 genomic RNA is long-lived after entry, suggesting that it avoids degradation by cellular nucleases. Moreover, we observed that SARS-CoV-2 replication is highly variable between cells, with only a small cell population displaying high burden of viral RNA. Unexpectedly, the B.1.1.7 variant, first identified in the UK, exhibits significantly slower replication kinetics than the Victoria strain, suggesting a novel mechanism contributing to its higher transmissibility with important clinical implications.

Democratising "Microscopi": a 3D printed automated XYZT fluorescence imaging system for teaching, outreach and fieldwork.

Commercial fluorescence microscope stands and fully automated XYZt fluorescence imaging systems are generally beyond the limited budgets available for teaching and outreach. We have addressed this problem by developing "Microscopi", an accessible, affordable, DIY automated imaging system that is built from 3D printed and commodity off-the-shelf hardware, including electro-mechanical, computer and optical components. Our design features automated sample navigation and image capture with a simple web-based graphical user interface, accessible with a tablet or other mobile device. The light path can easily be switched between different imaging modalities. The open source Python-based control software allows the hardware to be driven as an integrated imaging system. Furthermore, the microscope is fully customisable, which also enhances its value as a learning tool. Here, we describe the basic design and demonstrate imaging performance for a range of easily sourced specimens.

Microscope-Cockpit: Python-based bespoke microscopy for bio-medical science.

We have developed "Microscope-Cockpit" (Cockpit), a highly adaptable open source user-friendly Python-based Graphical User Interface (GUI) environment for precision control of both simple and elaborate bespoke microscope systems. The user environment allows next-generation near instantaneous navigation of the entire slide landscape for efficient selection of specimens of interest and automated acquisition without the use of eyepieces. Cockpit uses "Python-Microscope" (Microscope) for high-performance coordinated control of a wide range of hardware devices using open source software. Microscope also controls complex hardware devices such as deformable mirrors for aberration correction and spatial light modulators for structured illumination via abstracted device models. We demonstrate the advantages of the Cockpit platform using several bespoke microscopes, including a simple widefield system and a complex system with adaptive optics and structured illumination. A key strength of Cockpit is its use of Python, which means that any microscope built with Cockpit is ready for future customisation by simply adding new libraries, for example machine learning algorithms to enable automated microscopy decision making while imaging.

CytoCensus, mapping cell identity and division in tissues and organs using machine learning.

A major challenge in cell and developmental biology is the automated identification and quantitation of cells in complex multilayered tissues. We developed CytoCensus: an easily deployed implementation of supervised machine learning that extends convenient 2D 'point-and-click' user training to 3D detection of cells in challenging datasets with ill-defined cell boundaries. In tests on such datasets, CytoCensus outperforms other freely available image analysis software in accuracy and speed of cell detection. We used CytoCensus to count stem cells and their progeny, and to quantify individual cell divisions from time-lapse movies of explanted Drosophila larval brains, comparing wild-type and mutant phenotypes. We further illustrate the general utility and future potential of CytoCensus by analysing the 3D organisation of multiple cell classes in Zebrafish retinal organoids and cell distributions in mouse embryos. CytoCensus opens the possibility of straightforward and robust automated analysis of developmental phenotypes in complex tissues.

There are around 200 billion cells in the human brain that are generated by a small pool of rapidly dividing stem cells. For the brain to develop correctly, these stem cells must produce an appropriate number of each type of cell in the right place, at the right time. However, it remains unclear how individual stem cells in the brain know when and where to divide. To answer this question, Hailstone et al. studied the larvae of fruit flies, which use similar genes and mechanisms as humans to control brain development. This involved devising a new method for extracting the brains of developing fruit flies and keeping the intact tissue alive for up to 24 hours while continuously imaging individual cells in three dimensions. Manually tracking the division of each cell across multiple frames of a time-lapse is extremely time consuming. To tackle this problem, Hailstone et al. created a tool called CytoCensus, which uses machine learning to automatically identify stem cells from three-dimensional images and track their rate of division over time. Using the CytoCensus tool, Hailstone et al. identified a gene that controls the diverse rates at whichstem cells divide in the brain. Earlier this year some of the same researchers also published a study showing that this gene regulates a well-known cancer-related protein using an unconventional mechanism. CytoCensus was also able to detect cells in other developing tissues, including the embryos of mice. In the future, this tool could aid research into diseases that affect complex tissues, such as neurodegenerative disorders and cancer.

CryoSIM: super-resolution 3D structured illumination cryogenic fluorescence microscopy for correlated ultrastructural imaging.

Rapid cryopreservation of biological specimens is the gold standard for visualizing cellular structures in their true structural context. However, current commercial cryo-fluorescence microscopes are limited to low resolutions. To fill this gap, we have developed cryoSIM, a microscope for 3D super-resolution fluorescence cryo-imaging for correlation with cryo-electron microscopy or cryo-soft X-ray tomography. We provide the full instructions for replicating the instrument mostly from off-the-shelf components and accessible, user-friendly, open-source Python control software. Therefore, cryoSIM democratizes the ability to detect molecules using super-resolution fluorescence imaging of cryopreserved specimens for correlation with their cellular ultrastructure.

Testing Models of mRNA Localization Reveals Robustness Regulated by Reducing Transport between Cells.

Robust control of gene expression in both space and time is of central importance in the regulation of cellular processes and for multicellular development. However, the mechanisms by which robustness is achieved are generally not identified or well understood. For example, messenger RNA (mRNA) localization by molecular motor-driven transport is crucial for cell polarization in numerous contexts, but the regulatory mechanisms that enable this process to take place in the face of noise or significant perturbations are not fully understood. Here, we use a combined experimental-theoretical approach to characterize the robustness of gurken/transforming growth factor-α mRNA localization in Drosophila egg chambers, where the oocyte and 15 surrounding nurse cells are connected in a stereotypic network via intracellular bridges known as ring canals. We construct a mathematical model that encodes simplified descriptions of the range of steps involved in mRNA localization, including production and transport between and within cells until the final destination in the oocyte. Using Bayesian inference, we calibrate this model using quantitative single molecule fluorescence in situ hybridization data. By analyzing both the steady state and dynamic behaviors of the model, we provide estimates for the rates of different steps of the localization process as well as the extent of directional bias in transport through the ring canals. The model predicts that mRNA synthesis and transport must be tightly balanced to maintain robustness, a prediction that we tested experimentally using an overexpression mutant. Surprisingly, the overexpression mutant fails to display the anticipated degree of overaccumulation of mRNA in the oocyte predicted by the model. Through careful model-based analysis of quantitative data from the overexpression mutant, we show evidence of saturation of the transport of mRNA through ring canals. We conclude that this saturation engenders robustness of the localization process in the face of significant variation in the levels of mRNA synthesis.

IsoSense: frequency enhanced sensorless adaptive optics through structured illumination.

We present IsoSense, a wavefront sensing method that mitigates sample dependency in image-based sensorless adaptive optics applications in microscopy. Our method employs structured illumination to create additional high spatial frequencies in the image through custom illumination patterns. This improves the reliability of image quality metric calculations and enables sensorless wavefront measurement even in samples with sparse spatial frequency content. We demonstrate the feasibility of IsoSense for aberration correction in a deformable-mirror-based structured illumination super-resolution fluorescence microscope.

Accessibility to Peptidoglycan Is Important for the Recognition of Gram-Positive Bacteria in Drosophila.

In Drosophila, it is thought that peptidoglycan recognition proteins (PGRPs) SA and LC structurally discriminate between bacterial peptidoglycans with lysine (Lys) or diaminopimelic (DAP) acid, respectively, thus inducing differential antimicrobial transcription response. Here, we find that accessibility to PG at the cell wall plays a central role in immunity to infection. When wall teichoic acids (WTAs) are genetically removed from S. aureus (Lys type) and Bacillus subtilis (DAP type), thus increasing accessibility, the binding of both PGRPs to either bacterium is increased. PGRP-SA and -LC double mutant flies are more susceptible to infection with both WTA-less bacteria. In addition, WTA-less bacteria grow better in PGRP-SA/-LC double mutant flies. Finally, infection with WTA-less bacteria abolishes any differential activation of downstream antimicrobial transcription. Our results indicate that accessibility to cell wall PG is a major factor in PGRP-mediated immunity and may be the cause for discrimination between classes of pathogens.

Super-Resolution Single Molecule FISH at the Drosophila Neuromuscular Junction.

The lack of an effective, simple, and highly sensitive protocol for fluorescent in situ hybridization (FISH) at the Drosophila larval neuromuscular junction (NMJ) has hampered the study of mRNA biology. Here, we describe our modified single molecule FISH (smFISH) methods that work well in whole mount Drosophila NMJ preparations to quantify primary transcription and count individual cytoplasmic mRNA molecules in specimens while maintaining ultrastructural preservation. The smFISH method is suitable for high-throughput sample processing and 3D image acquisition using any conventional microscopy imaging modality and is compatible with the use of antibody colabeling and transgenic fluorescent protein tags in axons, glia, synapses, and muscle cells. These attributes make the method particularly amenable to super-resolution imaging. With 3D Structured Illumination Microscopy (3D-SIM), which increases spatial resolution by a factor of 2 in X, Y, and Z, we acquire super-resolution information about the distribution of single molecules of mRNA in relation to covisualized synaptic and cellular structures. Finally, we demonstrate the use of commercial and open source software for the quality control of single transcript expression analysis, 3D-SIM data acquisition and reconstruction as well as image archiving management and presentation. Our methods now allow the detailed mechanistic and functional analysis of sparse as well as abundant mRNAs at the NMJ in their appropriate cellular context.

Localized Translation of gurken/TGF-α mRNA during Axis Specification Is Controlled by Access to Orb/CPEB on Processing Bodies.

In Drosophila oocytes, gurken/TGF-α mRNA is essential for establishing the future embryonic axes. gurken remains translationally silent during transport from its point of synthesis in nurse cells to its final destination in the oocyte, where it associates with the edge of processing bodies. Here we show that, in nurse cells, gurken is kept translationally silent by the lack of sufficient Orb/CPEB, its translational activator. Processing bodies in nurse cells have a similar protein complement and ultrastructure to those in the oocyte, but they markedly less Orb and do not associate with gurken mRNA. Ectopic expression of Orb in nurse cells at levels similar to the wild-type oocyte dorso-anterior corner at mid-oogenesis is sufficient to cause gurken mRNA to associate with processing bodies and translate prematurely. We propose that controlling the spatial distribution of translational activators is a fundamental mechanism for regulating localized translation.

A single and rapid calcium wave at egg activation in Drosophila.

Activation is an essential process that accompanies fertilisation in all animals and heralds major cellular changes, most notably, resumption of the cell cycle. While activation involves wave-like oscillations in intracellular Ca(2+) concentration in mammals, ascidians and polychaete worms and a single Ca(2+) peak in fish and frogs, in insects, such as Drosophila, to date, it has not been shown what changes in intracellular Ca(2+) levels occur. Here, we utilise ratiometric imaging of Ca(2+) indicator dyes and genetically encoded Ca(2+) indicator proteins to identify and characterise a single, rapid, transient wave of Ca(2+) in the Drosophila egg at activation. Using genetic tools, physical manipulation and pharmacological treatments we demonstrate that the propagation of the Ca(2+) wave requires an intact actin cytoskeleton and an increase in intracellular Ca(2+) can be uncoupled from egg swelling, but not from progression of the cell cycle. We further show that mechanical pressure alone is not sufficient to initiate a Ca(2+) wave. We also find that processing bodies, sites of mRNA decay and translational regulation, become dispersed following the Ca(2+) transient. Based on this data we propose the following model for egg activation in Drosophila: exposure to lateral oviduct fluid initiates an increase in intracellular Ca(2+) at the egg posterior via osmotic swelling, possibly through mechano-sensitive Ca(2+) channels; a single Ca(2+) wave then propagates in an actin dependent manner; this Ca(2+) wave co-ordinates key developmental events including resumption of the cell cycle and initiation of translation of mRNAs such as bicoid.

Subcellular mRNA localisation at a glance.

mRNA localisation coupled to translational regulation provides an important means of dictating when and where proteins function in a variety of model systems. This mechanism is particularly relevant in polarised or migrating cells. Although many of the models for how this is achieved were first proposed over 20 years ago, some of the molecular details are still poorly understood. Nevertheless, advanced imaging, biochemical and computational approaches have started to shed light on the cis-acting localisation signals and trans-acting factors that dictate the final destination of localised transcripts. In this Cell Science at a Glance article and accompanying poster, we provide an overview of mRNA localisation, from transcription to degradation, focusing on the microtubule-dependent active transport and anchoring mechanism, which we will use to explain the general paradigm. However, it is clear that there are diverse ways in which mRNAs become localised and target protein expression, and we highlight some of the similarities and differences between these mechanisms.

Drosophila patterning is established by differential association of mRNAs with P bodies.

The primary embryonic axes in flies, frogs and fish are formed through translational regulation of localized transcripts before fertilization. In Drosophila melanogaster, the axes are established through the transport and translational regulation of gurken (grk) and bicoid (bcd) messenger RNA in the oocyte and embryo. Both transcripts are translationally silent while being localized within the oocyte along microtubules by cytoplasmic dynein. Once localized, grk is translated at the dorsoanterior of the oocyte to send a TGF-α signal to the overlying somatic cells. In contrast, bcd is translationally repressed in the oocyte until its activation in early embryos when it forms an anteroposterior morphogenetic gradient. How this differential translational regulation is achieved is not fully understood. Here, we address this question using ultrastructural analysis, super-resolution microscopy and live-cell imaging. We show that grk and bcd ribonucleoprotein (RNP) complexes associate with electron-dense bodies that lack ribosomes and contain translational repressors. These properties are characteristic of processing bodies (P bodies), which are considered to be regions of cytoplasm where decisions are made on the translation and degradation of mRNA. Endogenous grk mRNA forms dynamic RNP particles that become docked and translated at the periphery of P bodies, where we show that the translational activator Oo18 RNA-binding protein (Orb, a homologue of CEPB) and the anchoring factor Squid (Sqd) are also enriched. In contrast, an excess of grk mRNA becomes localized inside the P bodies, where endogenous bcd mRNA is localized and translationally repressed. Interestingly, bcd mRNA dissociates from P bodies in embryos following egg activation, when it is known to become translationally active. We propose a general principle of translational regulation during axis specification involving remodelling of transport RNPs and dynamic partitioning of different transcripts between the translationally active edge of P bodies and their silent core.

Preparing individual Drosophila egg chambers for live imaging.

Live cell imaging is an important technique applied to a number of Drosophila tissues used as models to investigate topics such as axis specification, cell differentiation and organogenesis (1). Correct preparation of the experimental samples is a crucial, often neglected, step. The goal of preparation is to ensure physiological relevance and to establish optimal imaging conditions. To maintain tissue viability, it is critical to avoid dehydration, hypoxia, overheating or medium deterioration (2). The Drosophila egg chamber is a well established system for examining questions relating, but not limited, to body patterning, mRNA localization and cytoskeletal organization (3,4). For early- and mid-stage egg chambers, mounting in halocarbon oil is good for survival in that it allows free diffusion of oxygen, prevents dehydration and hypoxia and has superb optical properties for microscopy. Imaging of fluorescent proteins is possible through the introduction of transgenes into the egg chamber or physical injection of labeled RNA, protein or antibodies (5-7). For example, addition of MS2 constructs to the genome of animals enables real time observation of mRNAs in the oocyte (8). These constructs allow for in vivo labeling of mRNA through utilization of the MS2 bacteriophage RNA stem loop interaction with its coat protein (9). Here, we present a protocol for the extraction of ovaries as well as isolating individual ovarioles and egg chambers from the female Drosophila. For a detailed description of Drosophila oogenesis see Allan C. Spradling (1993, reprinted 2009) (10).

A cell biologist's guide to high resolution imaging.

Fluorescence microscopy is particularly well suited to the study of cell biology, due to its noninvasive nature, high sensitivity detection of specific molecules, and high spatial and temporal resolution. In recent years, there has been an important transition from imaging the static distributions of molecules as a snapshot in time in fixed material to live-cell imaging of the dynamics of molecules in cells: in essence visualizing biochemical processes in living cells. Furthermore, in the last 5 years, there have been important advances in so-called "super-resolution" imaging methods that have overcome the resolution limits imposed by the diffraction of light in optical systems. Live-cell imaging is now beginning to deliver in unprecedented detail, bridging the resolution gap between electron microscopy and light microscopy. We discuss the various factors that limit the spatial and temporal resolution of microscopy and how to overcome them, how to best prepare specimens for high resolution imaging, and the choice of fluorochromes. We also summarize the pros and cons of the different super-resolution techniques and introduce some of the key data analysis tasks that a cell biologist employing high resolution microscopy is typically interested in.

Remodelling of cortical actin where lytic granules dock at natural killer cell immune synapses revealed by super-resolution microscopy.

Natural Killer (NK) cells are innate immune cells that secrete lytic granules to directly kill virus-infected or transformed cells across an immune synapse. However, a major gap in understanding this process is in establishing how lytic granules pass through the mesh of cortical actin known to underlie the NK cell membrane. Research has been hampered by the resolution of conventional light microscopy, which is too low to resolve cortical actin during lytic granule secretion. Here we use two high-resolution imaging techniques to probe the synaptic organisation of NK cell receptors and filamentous (F)-actin. A combination of optical tweezers and live cell confocal microscopy reveals that microclusters of NKG2D assemble into a ring-shaped structure at the centre of intercellular synapses, where Vav1 and Grb2 also accumulate. Within this ring-shaped organisation of NK cell proteins, lytic granules accumulate for secretion. Using 3D-structured illumination microscopy (3D-SIM) to gain super-resolution of ~100 nm, cortical actin was detected in a central region of the NK cell synapse irrespective of whether activating or inhibitory signals dominate. Strikingly, the periodicity of the cortical actin mesh increased in specific domains at the synapse when the NK cell was activated. Two-colour super-resolution imaging revealed that lytic granules docked precisely in these domains which were also proximal to where the microtubule-organising centre (MTOC) polarised. Together, these data demonstrate that remodelling of the cortical actin mesh occurs at the central region of the cytolytic NK cell immune synapse. This is likely to occur for other types of cell secretion and also emphasises the importance of emerging super-resolution imaging technology for revealing new biology.