Human SAGA is a 20-subunit transcriptional coactivator. Compared with yeast, metazoan SAGA uniquely incorporates a 150-kDa splicing-factor module (SPL), also present in U2 small nuclear ribonucleoprotein (U2snRNP). Metazoan gene duplication further specialized shared TFIID/SAGA subunits into SAGA-specific paralogs (TAF5L and TAF6L), but the functional consequences of this divergence are unknown. We report the structure of endogenous human SAGA purified via an affinity ligand from cells that were not disturbed by any genomic engineering tools. Our work reveals the high-resolution structure of SPL and the TAF6L HEAT repeat domain that provides the SPL with a docking surface. We elucidate how SPL and the HEAT repeats are incorporated into SAGA. We identify major structural differences between TAF6L/TAF5L and their canonical paralogs that enable SPL accommodation. SPL engages SAGA through a substantially smaller interface than in U2snRNP, despite sharing a deeply inserted helical motif. The seemingly weaker interaction of SPL with SAGA raises the possibility that SAGA relays this module to the splicing machinery.

Deep learning methods using convolutional neural networks are very effective for automatic image segmentation tasks with no exception for cellular electron micrographs. However, the lack of dedicated easy-to-use tools largely reduces the widespread use of these techniques. Here we present DeepSCEM, a straightforward tool for fast and efficient segmentation of cellular electron microscopy images using deep learning with a special focus on efficient and user-friendly generation and training of models for organelle segmentation.

Chromatin structure is a key regulator of DNA transcription, replication and repair1. In humans, the TIP60-EP400 complex (TIP60-C) is a 20-subunit assembly that affects chromatin structure through two enzymatic activities: ATP-dependent exchange of histone H2A-H2B for H2A.Z-H2B, and histone acetylation. In yeast, however, these activities are performed by two independent complexes-SWR1 and NuA4, respectively2,3. How the activities of the two complexes are merged into one supercomplex in humans, and what this association entails for the structure and mechanism of the proteins and their recruitment to chromatin, are unknown. Here we describe the structure of the endogenous human TIP60-C. We find a three-lobed architecture composed of SWR1-like (SWR1L) and NuA4-like (NuA4L) parts, which associate with a TRRAP activator-binding module. The huge EP400 subunit contains the ATPase motor, traverses the junction between SWR1L and NuA4L twice and constitutes the scaffold of the three-lobed architecture. NuA4L is completely rearranged compared with its yeast counterpart. TRRAP is flexibly tethered to NuA4L-in stark contrast to its robust connection to the completely opposite side of NuA4 in yeast4-7. A modelled nucleosome bound to SWR1L, supported by tests of TIP60-C activity, suggests that some aspects of the histone exchange mechanism diverge from what is seen in yeast8,9. Furthermore, a fixed actin module (as opposed to the mobile actin subcomplex in SWR1; ref. 8), the flexibility of TRRAP and the weak effect of extranucleosomal DNA on exchange activity lead to a different, activator-based mode of enlisting TIP60-C to chromatin.

Sirtuin 6 (SIRT6) is an NAD + dependent histone H3 deacetylase that is prominently found associated with chromatin, attenuates transcriptionally active promoters and regulates DNA repair, metabolic homeostasis and lifespan. Unlike other sirtuins, it has low affinity to free histone tails but demonstrates strong binding to nucleosomes. It is poorly understood how SIRT6 docking on nucleosomes stimulates its histone deacetylation activity. Here we present the structure of human SIRT6 bound to a nucleosome determined by cryogenic electron microscopy. The zinc finger domain of SIRT6 associates tightly with the acidic patch of the nucleosome through multiple arginine anchors. The Rossmann fold domain binds to the terminus of the looser DNA half of the nucleosome, detaching two turns of the DNA from the histone octamer and placing the NAD + binding pocket close to the DNA exit site. This domain shows flexibility with respect to the fixed zinc finger and moves with, but also relative to, the unwrapped DNA terminus. We apply molecular dynamics simulations of the histone tails in the nucleosome to show that in this mode of interaction, the active site of SIRT6 is perfectly poised to catalyze deacetylation of the H3 histone tail and that the partial unwrapping of the DNA allows even lysines close to the H3 core to reach the enzyme.

Max-tree (or component-tree) is a hierarchical representation which associates to a scalar image a descriptive data structure induced by the inclusion relation between the binary components obtained at successive level-sets. Various attributes related to these binary components can be computed and stored into the tree. Max-trees have been involved in many applications, enabling to perform attribute filtering in an efficient algorithmic way. Since the resulting images do not contain any new contour, these kind of filters are called connected operators . In this paper, we propose to rely on max-trees and attribute filters to enrich the input of a convolutional neural network (CNN) to improve a task of segmentation. More precisely, two approaches are considered: a first approach in which images are preprocessed using attribute filters and a second approach in which maps of attributes relying on max-trees are computed. Based on these two different approaches, the resulting maps are used as additional input in a standard CNN in a context of semantic segmentation. We propose to compare different attributes and nodes selection strategies and to experiment their usage on a practical problem: the segmentation of the mitochondria and endoplasmic-reticulum in Focused Ion Beam milling combined with Scanning Electron Microscopy (FIB-SEM) images. We provide original images, annotations, source code and a documentation to reproduce the experimentation results.

Advances in cryo-electron microscopy (EM) enable imaging of protein assemblies within mammalian cells in a near native state when samples are preserved by cryogenic vitrification. To accompany this progress, specialized EM labelling protocols must be developed. Gold nanoparticles (AuNPs) of 2 nm are synthesized and functionalized to bind selected intracellular targets inside living human cells and to be detected in vitreous sections. As a proof of concept, thioaminobenzoate-, thionitrobenzoate-coordinated gold nanoparticles are functionalized on their surface with SV40 Nuclear Localization Signal (NLS)-containing peptides and 2 kDa polyethyleneglycols (PEG) by thiolate exchange to target the importin-mediated nuclear machinery and facilitate cytosolic diffusion by shielding the AuNP surface from non-specific binding to cell components, respectively. After delivery by electroporation into the cytoplasm of living human cells, the PEG-coated AuNPs diffuse freely in the cytoplasm but do not enter the nucleus. Incorporation of NLS within the PEG coverage promotes a quick nuclear import of the nanoparticles in relation to the density of NLS onto the AuNPs. Cryo-EM of vitreous cell sections demonstrate the presence of 2 nm AuNPs as single entities in the nucleus. Biofunctionalized AuNPs combined with live-cell electroporation procedures are thus potent labeling tools for the identification of macromolecules in cellular cryo-EM.

Gene transcription initiation is a highly regulated process in which Pol II and general transcription factors assemble into a pre-initiation complex. Structural studies of yeast and human initiation complexes shed light on the role of the first nucleosome flanking gene promoters in controlling the transcription machinery.

Despite advances in Electron Microscopy (EM) that enable to image protein assemblies within vitreous sections of cells at nearly atomic resolution, labelling is still necessary to locate small proteins or rare complexes. Gold immunolabelling has been used for decades to localise specific proteins within cellular sections. However, current gold particle-antibody conjugates are not built with enough chemical precision to match the current resolution offered by cryo-EM methodology. Furthermore, as a close to native specimen state can only be achieved by strict preservation of a frozen hydrated state, it is required to deliver gold labelling agents into living cells prior to their vitrification. Several 1.4 nm gold nanoparticle-antibody conjugates were synthesised. Their abilities to bind to and label their corresponding epitopes within living cells after cytosolic delivery by electroporation are documented here.

Histone acetylation regulates most DNA transactions and is dynamically controlled by highly conserved enzymes. The only essential histone acetyltransferase (HAT) in yeast, Esa1, is part of the 1-MDa NuA4 complex, which plays pivotal roles in both transcription and DNA-damage repair. NuA4 has the unique capacity to acetylate histone targets located several nucleosomes away from its recruitment site. Neither the molecular mechanism of this activity nor its physiological importance are known. Here we report the structure of the Pichia pastoris NuA4 complex, with its core resolved at 3.4-Å resolution. Three subunits, Epl1, Eaf1 and Swc4, intertwine to form a stable platform that coordinates all other modules. The HAT module is firmly anchored into the core while retaining the ability to stretch out over a long distance. We provide structural, biochemical and genetic evidence that an unfolded linker region of the Epl1 subunit is critical for this long-range activity. Specifically, shortening the Epl1 linker causes severe growth defects and reduced H4 acetylation levels over broad chromatin regions in fission yeast. Our work lays the foundations for a mechanistic understanding of NuA4’s regulatory role and elucidates how its essential long-range activity is attained.

Abstract Sirtuin 6 (SIRT6) is a NAD+ dependent histone H3 deacetylase that is prominently found associated with chromatin, attenuates transcriptionally active promoters and regulates DNA repair, metabolic homeostasis and life-span. Unlike other sirtuins it has low affinity to free histone tail but demonstrates strong binding to nucleosomes. It is poorly understood how SIRT6 docking on nucleosomes stimulates its histone de-acetylation activity. Here we present the structure of human SIRT6 bound to a nucleosome determined by cryogenic electron microscopy. The zinc finger domain of SIRT6 associates tightly with the acidic patch of the nucleosome through multiple arginine anchors. The Rossmann fold domain binds to the terminus of the looser DNA half of the nucleosome, detaching two turns of the DNA from the histone octamer and placing the NAD+ binding pocket close to the DNA exit site. This domain shows flexibility with respect to the fixed zinc finger and moves with, but also relative to, the unwrapped DNA terminus. We apply molecular dynamics simulations of the histone tails in the nucleosome to show that in this mode of interaction the active site of SIRT6 is perfectly poised to catalyze deacetylation of the H3 histone tail and that the partial unwrapping of the DNA allows even lysines close to the H3 core to reach the enzyme.

Electron microscopy imaging techniques allow biologists to obtain nanoscale slice volumes of cells in which biological structures can be statistically analyzed under the condition that the tedious task of segmentation can be automated. If a learned segmentation has good results on images obtained under the same acquisition process, variations in this acquisition can drastically affect the final image’s features and can lead to a drop of segmentation performances. To tackle this problem, we propose an Unsupervised Domain Adaptation (UDA) framework based on Batch Normalization. Our method tested on mitochondria segmentation shows better results than state-of-the art methods and uses few computation resources, allowing a near real-time use. Furthermore, we propose a metric based on the Wasserstein distance to evaluate the effect of this normalization. We empirically show that this measure is log-linearly correlated with the drop of performances of the adapted segmentation and sheds light on the gain obtained by our UDA framework.

A full 3D analysis of the hierarchical porosity in Coscinodiscus sp. diatom structures was carried out by using a multiscale approach that combines three advanced volumetric imaging techniques with resolutions and fields of view covering all the porous characteristics of such complex architectures: electron tomography, "slice and view" approach that uses a dual-beam microscope (FIB-SEM), and array tomography consisting of serial imaging of ultrathin specimen sections. This multiscale approach allowed the whole porosity network to be quantified and provided an unprecedented structural insight into these natural nanostructured materials with internal organization ranging from micrometer to nanometer. The analysed species is made of several nested layers with different pore sizes, shapes and connectivities and characterized by the presence of interconnected pores structured in various ways. The first evidence of the presence of a nanometric porosity made of ellipsoidal pores in the siliceous diatom frustules is also provided.

Cancer is a multi-step disease where an initial tumour progresses through critical steps shaping, in most cases, life-threatening secondary foci called metastases. The oncogenic cascade involves genetic, epigenetic, signalling pathways, intracellular trafficking and/or metabolic alterations within cancer cells. In addition, pre-malignant and malignant cells orchestrate complex and dynamic interactions with non-malignant cells and acellular matricial components or secreted factors within the tumour microenvironment that is instrumental in the progression of the disease. As our aptitude to effectively treat cancer mostly depends on our ability to decipher, properly diagnose and impede cancer progression and metastasis formation, full characterisation of molecular complexes and cellular processes at play along the metastasis cascade is crucial. For many years, the scientific community lacked adapted imaging and molecular technologies to accurately dissect, at the highest resolution possible, tumour and stromal cells behaviour within their natural microenvironment. In that context, the NANOTUMOR consortium is a French national multi-disciplinary workforce which aims at a providing a multi-scale characterisation of the oncogenic cascade, from the atomic level to the dynamic organisation of the cell in response to genetic mutations, environmental changes or epigenetic modifications. Ultimately, this program aims at identifying new therapeutic targets using innovative drug design.

Focused Ion Beam milling combined with Scanning Electron Microscopy (FIB-SEM) technique is an electron microscopy imaging method that offers the possibility of acquiring 3D isotropic images of biological structures at the nanometric scale. Automated image segmentation is required for morphological analysis of huge image stacks and to save time consuming manual intervention. Current methods are either specific to data and organelles or lack accuracy. We propose a robust multi-class semantic segmentation method for FIBSEM images, based on deep neural networks. We evaluate and compare our proposed method on two FIB-SEM images, for the segmentation of mitochondria, cell membrane and endoplasmic reticulum. We achieve results close to inter-expert variability.

The transcription of eukaryotic protein genes is controlled by a plethora of proteins which act together in multi-component complexes to facilitate the DNA dependent RNA polymerase II (Pol II) enzyme to bind to the transcription start site and to generate messenger RNA from the gene's coding sequence. The protein that guides the transcription machinery to the exact transcription start site is called TATA-box Binding Protein, or TBP. TBP is part of two large protein complexes involved in Pol II transcription, TFIID and SAGA. The two complexes share several subunits implicated in the interaction with TBP and contain proteins with structural elements highly homologous to nucleosomal histones. Despite the intensive study of transcription initiation, the mode of interaction of TBP with these complexes and its release upon DNA binding was elusive. In this study we demonstrate the quasi-atomic model of SAGA in complex with TBP. The structure reveals the intricate network of interactions that coordinate the different functional domains of SAGA and resolves a deformed octamer of histone-fold domains at the core of SAGA. This deformed octamer is precisely tuned to establish a peripheral site for TBP binding, where it is protected by steric hindrance against the binding of spurious DNA. Complementary biochemical analysis points to a mechanism for TBP delivery and release from SAGA that requires the general transcription factor TFIIA and whose efficiency correlates with the affinity of DNA to TBP.As the TBP binding machinery is highly similar in TFIID and SAGA, we demonstrated a universal mechanism of how TBP is delivered to gene promoters during transcription initiation.

The transcription of eukaryotic protein genes is controlled by a plethora of proteins which act together in multi-component complexes to facilitate the DNA dependent RNA polymerase II (Pol II) enzyme to bind to the transcription start site and to generate messenger RNA from the gene’s coding sequence. The protein that guides the transcription machinery to the exact transcription start site is called TATA-box Binding Protein, or TBP. TBP is part of two large protein complexes involved in Pol II transcription, TFIID and SAGA. The two complexes share several subunits implicated in the interaction with TBP and contain proteins with structural elements highly homologous to nucleosomal histones. Despite the intensive study of transcription initiation, the mode of interaction of TBP with these complexes and its release upon DNA binding was elusive. In this study we demonstrate the quasi-atomic model of SAGA in complex with TBP. The structure reveals the intricate network of interactions that coordinate the different functional domains of SAGA and resolves a deformed octamer of histone-fold domains at the core of SAGA. This deformed octamer is precisely tuned to establish a peripheral site for TBP binding, where it is protected by steric hindrance against the binding of spurious DNA. Complementary biochemical analysis points to a mechanism for TBP delivery and release from SAGA that requires the general transcription factor TFIIA and whose efficiency correlates with the affinity of DNA to TBP. ∗Corresponding author. ISSN (electronic) : 1768-3238 https://comptes-rendus.academie-sciences.fr/biologies/ 248 Gabor Papai et al. As the TBP binding machinery is highly similar in TFIID and SAGA, we demonstrated a universal mechanism of how TBP is delivered to gene promoters during transcription initiation. Résumé. La transcription des gènes des protéines eucaryotes est contrôlée par une pléthore de protéines agissant de concert sous forme de complexes multi-composants pour faciliter la liaison de l’enzyme ARN polymérase II ADN-dépendante (Pol II) au site d’initiation de la transcription et pour générer un ARN messager à partir de la séquence codante du gène. La protéine qui guide la machinerie de transcription vers le site d’initiation de la transcription est appelée protéine de liaison à la boîte TATA, ou TBP. TBP fait partie de deux complexes protéiques impliqués dans la transcription par la Pol II, TFIID et SAGA. Les deux complexes partagent plusieurs sous-unités impliquées dans l’interaction avec TBP et comportent des protéines présentant des éléments structuraux hautement homologues aux histones nucléosomiques. Malgré l’étude intensive de l’initiation de la transcription, le mode d’interaction de TBP avec ces complexes ainsi que sa libération lors de sa liaison de l’ADN étaient évasifs. Dans cette étude, nous avons déterminé un modèle quasi-atomique de SAGA en complexe avec TBP. La structure révèle le réseau d’interactions qui coordonnent les différents domaines fonctionnels de SAGA et résout un octamère déformé des domaines homologues aux histones au cœur de SAGA. Cet octamère déformé est précisément adapté pour établir un site périphérique de liaison à TBP, où ce dernier est protégé par une inhibition stérique contre la fixation d’un ADN parasite. L’analyse biochimique complémentaire a mis en évidence un mécanisme de libération de TBP de SAGA qui nécessite le facteur de transcription général TFIIA et dont l’efficacité corrèle avec l’affinité de l’ADN pour TBP. Comme le mécanisme de liaison de TBP est très similaire dans TFIID et SAGA, nous avons mis en évidence un mécanisme universel décrivant la manière dont TBP est délivré aux promoteurs de gènes lors de l’initiation de la transcription.