Optics clustered to output unique solutions: a multi-laser facility for combined single molecule and ensemble microscopy

Drug-resistant EGFR mutations promote lung cancer by stabilizing interfaces in ligand-free kinase-active EGFR oligomers

The Epidermal Growth Factor Receptor (EGFR) is frequently found to be mutated in non-small cell lung cancer. Oncogenic EGFR has been successfully targeted by tyrosine kinase inhibitors, but acquired drug resistance eventually overcomes the efficacy of these treatments. Attempts to surmount this therapeutic challenge are hindered by a poor understanding of how and why cancer mutations specifically amplify ligand-independent EGFR auto-phosphorylation signals to enhance cell survival and how this amplification is related to ligand-dependent cell proliferation. Here we show that drug-resistant EGFR mutations manipulate the assembly of ligand-free, kinase-active oligomers to promote and stabilize the assembly of oligomer-obligate active dimer sub-units and circumvent the need for ligand binding. We reveal the structure and assembly mechanisms of these ligand-free, kinase-active oligomers, uncovering oncogenic functions for hitherto orphan transmembrane and kinase interfaces, and for the ectodomain tethered conformation of EGFR. Importantly, we find that the active dimer sub-units within ligand-free oligomers are the high affinity binding sites competent to bind physiological ligand concentrations and thus drive tumor growth, revealing a link with tumor proliferation. Our findings provide a framework for future drug discovery directed at tackling oncogenic EGFR mutations by disabling oligomer-assembling interactions.

A brief history of the octopus imaging facility to celebrate its 10th anniversary

Octopus (Optics Clustered to OutPut Unique Solutions) celebrated in June 2020 its 10th birthday. Based at Harwell, near Oxford, Octopus is an open access, peer reviewed, national imaging facility that offers successful U.K. applicants supported access to single molecule imaging, confocal microscopy, several flavours of superresolution imaging, light sheet microscopy, optical trapping and cryoscanning electron microscopy. Managed by a multidisciplinary team, Octopus has so far assisted >100 groups of U.K. and international researchers. Cross-fertilisation across fields proved to be a strong propeller of success underpinned by combining access to top-end instrumentation with a strong programme of imaging hardware and software developments. How Octopus was born, and highlights of the multidisciplinary output produced during its 10-year journey are reviewed below, with the aim of celebrating a myriad of collaborations with the U.K. scientific community, and reflecting on their scientific and societal impact.

Machine learning and big scientific data

This paper reviews some of the challenges posed by the huge growth of experimental data generated by the new generation of large-scale experiments at UK national facilities at the Rutherford Appleton Laboratory (RAL) site at Harwell near Oxford. Such 'Big Scientific Data' comes from the Diamond Light Source and Electron Microscopy Facilities, the ISIS Neutron and Muon Facility and the UK's Central Laser Facility. Increasingly, scientists are now required to use advanced machine learning and other AI technologies both to automate parts of the data pipeline and to help find new scientific discoveries in the analysis of their data. For commercially important applications, such as object recognition, natural language processing and automatic translation, deep learning has made dramatic breakthroughs. Google's DeepMind has now used the deep learning technology to develop their AlphaFold tool to make predictions for protein folding. Remarkably, it has been able to achieve some spectacular results for this specific scientific problem. Can deep learning be similarly transformative for other scientific problems? After a brief review of some initial applications of machine learning at the RAL, we focus on challenges and opportunities for AI in advancing materials science. Finally, we discuss the importance of developing some realistic machine learning benchmarks using Big Scientific Data coming from several different scientific domains. We conclude with some initial examples of our 'scientific machine learning' benchmark suite and of the research challenges these benchmarks will enable. This article is part of a discussion meeting issue 'Numerical algorithms for high-performance computational science'.

A global sampler of single particle tracking solutions for single molecule microscopy

The dependence on model-fitting to evaluate particle trajectories makes it difficult for single particle tracking (SPT) to resolve the heterogeneous molecular motions typical of cells. We present here a global spatiotemporal sampler for SPT solutions using a Metropolis-Hastings algorithm. The sampler does not find just the most likely solution but also assesses its likelihood and presents alternative solutions. This enables the estimation of the tracking error. Furthermore the algorithm samples the parameters that govern the tracking process and therefore does not require any tweaking by the user. We demonstrate the algorithm on synthetic and single molecule data sets. Metrics for the comparison of SPT are generalised to be applied to a SPT sampler. We illustrate using the example of the diffusion coefficient how the distribution of the tracking solutions can be propagated into a distribution of derived quantities. We also discuss the major challenges that are posed by the realisation of a SPT sampler.

Hydrophobic fluorescent probes introduce artifacts into single molecule tracking experiments due to non-specific binding

Single-molecule techniques are powerful tools to investigate the structure and dynamics of macromolecular complexes; however, data quality can suffer because of weak specific signal, background noise and dye bleaching and blinking. It is less well-known, but equally important, that non-specific binding of probe to substrates results in a large number of immobile fluorescent molecules, introducing significant artifacts in live cell experiments. Following from our previous work in which we investigated glass coating substrates and demonstrated that the main contribution to this non-specific probe adhesion comes from the dye, we carried out a systematic investigation of how different dye chemistries influence the behaviour of spectrally similar fluorescent probes. Single-molecule brightness, bleaching and probe mobility on the surface of live breast cancer cells cultured on a non-adhesive substrate were assessed for anti-EGFR affibody conjugates with 14 different dyes from 5 different manufacturers, belonging to 3 spectrally homogeneous bands (491 nm, 561 nm and 638 nm laser lines excitation). Our results indicate that, as well as influencing their photophysical properties, dye chemistry has a strong influence on the propensity of dye-protein conjugates to adhere non-specifically to the substrate. In particular, hydrophobicity has a strong influence on interactions with the substrate, with hydrophobic dyes showing much greater levels of binding. Crucially, high levels of non-specific substrate binding result in calculated diffusion coefficients significantly lower than the true values. We conclude that the physic-chemical properties of the dyes should be considered carefully when planning single-molecule experiments. Favourable dye characteristics such as photostability and brightness can be offset by the propensity of a conjugate for non-specific adhesion.

A stochastic model for electron multiplication charge-coupled devices--from theory to practice

Electron multiplication charge-coupled devices (EMCCD) are widely used for photon counting experiments and measurements of low intensity light sources, and are extensively employed in biological fluorescence imaging applications. These devices have a complex statistical behaviour that is often not fully considered in the analysis of EMCCD data. Robust and optimal analysis of EMCCD images requires an understanding of their noise properties, in particular to exploit fully the advantages of Bayesian and maximum-likelihood analysis techniques, whose value is increasingly recognised in biological imaging for obtaining robust quantitative measurements from challenging data. To improve our own EMCCD analysis and as an effort to aid that of the wider bioimaging community, we present, explain and discuss a detailed physical model for EMCCD noise properties, giving a likelihood function for image counts in each pixel for a given incident intensity, and we explain how to measure the parameters for this model from various calibration images.

Single molecule fluorescence detection and tracking in mammalian cells: the state-of-the-art and future perspectives

Insights from single-molecule tracking in mammalian cells have the potential to greatly contribute to our understanding of the dynamic behavior of many protein families and networks which are key therapeutic targets of the pharmaceutical industry. This is particularly so at the plasma membrane, where the method has begun to elucidate the mechanisms governing the molecular interactions that underpin many fundamental processes within the cell, including signal transduction, receptor recognition, cell-cell adhesion, etc. However, despite much progress, single-molecule tracking faces challenges in mammalian samples that hinder its general application in the biomedical sciences. Much work has recently focused on improving the methods for fluorescent tagging of target molecules, detection and localization of tagged molecules, which appear as diffraction-limited spots in charge-coupled device (CCD) images, and objectively establishing the correspondence between moving particles in a sequence of image frames to follow their diffusive behavior. In this review we outline the state-of-the-art in the field and discuss the advantages and limitations of the methods available in the context of specific applications, aiming at helping researchers unfamiliar with single molecules methods to plan out their experiments.

A systematic investigation of differential effects of cell culture substrates on the extent of artifacts in single-molecule tracking

Single-molecule techniques are being increasingly applied to biomedical investigation, notwithstanding the numerous challenges they pose in terms of signal-to-noise ratio issues. Non-specific binding of probes to glass substrates, in particular, can produce experimental artifacts due to spurious molecules on glass, which can be particularly deleterious in live-cell tracking experiments. In order to resolve the issue of non-specific probe binding to substrates, we performed systematic testing of a range of available surface coatings, using three different proteins, and then extended our assessment to the ability of these coatings to foster cell growth and retain non-adhesive properties. Linear PEG, a passivating agent commonly used both in immobilized-molecule single-molecule techniques and in tissue engineering, is able to both successfully repel non-specific adhesion of fluorescent probes and to foster cell growth when functionalized with appropriate adhesive peptides. Linear PEG treatment results in a significant reduction of tracking artifacts in EGFR tracking with Affibody ligands on a cell line expressing EGFR-eGFP. The findings reported herein could be beneficial to a large number of experimental situations where single-molecule or single-particle precision is required.

Cell wall constrains lateral diffusion of plant plasma-membrane proteins

A cell membrane can be considered a liquid-phase plane in which lipids and proteins theoretically are free to diffuse. Numerous reports, however, describe retarded diffusion of membrane proteins in animal cells. This anomalous diffusion results from a combination of structuring factors including protein-protein interactions, cytoskeleton corralling, and lipid organization into microdomains. In plant cells, plasma-membrane (PM) proteins have been described as relatively immobile, but the control mechanisms that structure the PM have not been studied. Here, we use fluorescence recovery after photobleaching to estimate mobility of a set of minimal PM proteins. These proteins consist only of a PM-anchoring domain fused to a fluorescent protein, but their mobilities remained limited, as is the case for many full-length proteins. Neither the cytoskeleton nor membrane microdomain structure was involved in constraining the diffusion of these proteins. The cell wall, however, was shown to have a crucial role in immobilizing PM proteins. In addition, by single-molecule fluorescence imaging we confirmed that the pattern of cellulose deposition in the cell wall affects the trajectory and speed of PM protein diffusion. Regulation of PM protein dynamics by the plant cell wall can be interpreted as a mechanism for regulating protein interactions in processes such as trafficking and signal transduction.

Cell wall constrains lateral diffusion of plant plasma-membrane proteins

A cell membrane can be considered a liquid-phase plane in which lipids and proteins theoretically are free to diffuse. Numerous reports, however, describe retarded diffusion of membrane proteins in animal cells. This anomalous diffusion results from a combination of structuring factors including protein-protein interactions, cytoskeleton corralling, and lipid organization into microdomains. In plant cells, plasma-membrane (PM) proteins have been described as relatively immobile, but the control mechanisms that structure the PM have not been studied. Here, we use fluorescence recovery after photobleaching to estimate mobility of a set of minimal PM proteins. These proteins consist only of a PM-anchoring domain fused to a fluorescent protein, but their mobilities remained limited, as is the case for many full-length proteins. Neither the cytoskeleton nor membrane microdomain structure was involved in constraining the diffusion of these proteins. The cell wall, however, was shown to have a crucial role in immobilizing PM proteins. In addition, by single-molecule fluorescence imaging we confirmed that the pattern of cellulose deposition in the cell wall affects the trajectory and speed of PM protein diffusion. Regulation of PM protein dynamics by the plant cell wall can be interpreted as a mechanism for regulating protein interactions in processes such as trafficking and signal transduction.