Search

Chromosomes of all species studied so far display a variety of higher-order organizational features such as domains, loops, or compartments. Many of these structures have been characterized from the genome-wide contact maps generated by chromosome conformation capture approaches (Hi-C, ChIA-PET,…). Indeed, DNA 3D structures translate as distinct patterns visible on these maps. We developed Chromosight, an algorithm based on computer vision approaches that automatically detect and quantify any type of pattern in contact data. Chromosight detects 3 times as many patterns as existing programs, while being faster and fit to any genome, including small, compact ones. Chromosight is user-friendly and can be extended to user-provided patterns. We validated the program by applying it to a variety of chromosomal structures found in mammals. Code and documentation: https://github.com/koszullab/chromosight

International audience

Two proximity ligation approaches are used to probe the gut’s phage-bacteria infection network using Meta3C and GRAAL scaffolding., The biochemical activities of microbial communities, or microbiomes, are essential parts of environmental and animal ecosystems. The dynamics, balance, and effects of these communities are strongly influenced by phages present in the population. Being able to characterize bacterium-phage relationships is therefore essential to investigate these ecosystems to the full extent of their complexity. However, this task is currently limited by (i) the ability to characterize complete bacterial and viral genomes from a complex mix of species and (ii) the difficulty to assign phage sequences to their bacterial hosts. We show that both limitations can be circumvented using meta3C, an experimental and computational approach that exploits the physical contacts between DNA molecules to infer their proximity. In a single experiment, dozens of bacterial and phage genomes present in a complex mouse gut microbiota were assembled and scaffolded de novo. The phage genomes were then assigned to their putative bacterial hosts according to the physical contacts between the different DNA molecules, opening new perspectives for a comprehensive picture of the genomic structure of the gut flora. Therefore, this work holds far-reaching implications for human health studies aiming to bridge the virome to the microbiome.

International audience; Eukaryotic chromosomes undergo movements that are involved in the regulation of functional processes such as DNA repair. To better understand the origin of these movements, we used fluorescence microscopy, image analysis and chromosome conformation capture to quantify the actin contribution to chromosome movements and interactions in budding yeast. We show that both the cytoskeletal and nuclear actin drive local chromosome movements, independently of Csm4, a putative LINC protein. Inhibition of actin polymerization reduces subtelomere dynamics, resulting in more confined territories and enrichment in subtelomeric contacts. Artificial tethering of actin to nuclear pores increased both nuclear pore complex (NPC) and subtelomere motion. Chromosome loci that were positioned away from telomeres exhibited reduced motion in the presence of an actin polymerization inhibitor but were unaffected by the lack of Csm4. We further show that actin was required for locus mobility that was induced by targeting the chromatin-remodeling protein Ino80. Correlated with this, DNA repair by homologous recombination was less efficient. Overall, interphase chromosome dynamics are modulated by the additive effects of cytoskeletal actin through forces mediated by the nuclear envelope and nuclear actin, probably through the function of actin in chromatin-remodeling complexes.

International audience; Genome-wide derivatives of the chromosome conformation capture (3C) technique are now well-established approaches to study the multiscale average organization of chromosomes from bacteria to mammals. However, the experimental parameters of the protocol have to be optimized for different species, and the downstream experimental products (i.e., pair-end sequences) are influenced by these parameters. Here, we describe a complete pipeline to generate 3C-seq libraries and compute chromosomal contact maps of yeast species.

Position and firing time of each replication origin, by chromosome. Coordinates are given according to Genolevures sequence release 10 September 2008 ( www.genolevures.org ). Bona fide origins are in red (see text). (DOC 154 kb)

ARS library. Number of ARSs (y axis) for each coverage (x axis) in the ARS library, before transformation. Coverage was determined by Illumina sequencing (see “Methods”). The average coverage of the library was 9×. Each bar represents 2× coverage. (PDF 98 kb)

Positions, coverages, and fitnesses of ARSs. (XLS 69 kb)

SIR-mediated telomere clustering drives chromosome conformation in the dense fraction of SP cells. a Mean contacts frequencies between 100-kb centromeres windows in G1 (blue) and G0 quiescent cells (red). Black and green curves: contacts between 100-kb segments randomly sampled in both conditions, to illustrate the absence of coverage biases after normalization. b Chromosome organization of WT and sir3∆ quiescent cells (the cryptic mating type locus HML was deleted to prevent pseudo-diploid effect). ii) Normalized contact matrix obtained for hml∆* (left) and hml∆ sir3∆ (right) cells. Color scale: contact frequencies from rare (white) to frequent (dark blue). Red arrowheads: centromeres contacts; green and yellow arrowheads: telomere–telomere contacts in hml∆ and hml∆ sir3∆ G0 cells, respectively. The 3D representations of the hml∆ and hml∆ sir3∆ matrices are represented next to the contact maps. Each chromosome is represented as a chain of beads (1 bead = 20 kb), with color code reflecting the chromosome arm lengths, from short (blue) to long (red) arms. Yellow beads: subtelomeric regions; black beads: centromeres; purple beads: boundaries of the rDNA cluster. c Contact maps of W303 strain during exponentially growth (EXPO, left) and quiescence (G0, right). Red arrowheads: centromere clustering; green and yellow arrowheads: telomere–telomere contacts of two chromosomes (XIII and XV) in expo and G0 cells, respectively. Because of the low sequencing coverage and quality, the signal is not as strong as for data in Fig. 3 and the bins are larger (1 vector: 80 DpnII RFs). d Quantification of colocalization of 30-kb telomeric regions (red dots) compared with the distribution of the colocalization scores (box plot, two standard deviations) computed for 1000 random sets of 32 windows of 30 kb in the genome (excluding centromeric regions). The colocalization score is normalized by the sequencing depth for each dataset.

Strains used in this study. (DOCX 22 kb)

Mechanism driving telomere clustering in long-lived SP cells. a Western blot against Sir3 and H2A on crude extracts from exponential, respiratory, or stationary cultures of a wild-type (WT) strain (yAT1684). b Sir3 spreading at yeast subtelomeres in cells from an exponentially growing culture (Fermentation) or in cells isolated from the dense fraction of a SP culture (Stationary HD). ChIP-chip profiles (Sir3 enrichment Z score) correspond to the mean of two independent experiments. Pearson correlation between conditions is 0.95. Sir3 spreading at TELVIR was confirmed in independent experiments by ChIP-quantitative PCR for both conditions (not shown). Each panel spans the first 30 kb from each telomere and the heading color for each panel indicates the middle repeat element content of the corresponding telomere: Y’ XCR XCS (beige), Y’ XCS (green), XCS (red), or XCR XCS (blue). Each dot represents a data point and lines are drawn for visual purposes. c Quiescent sir3∆ cells are as thermotolerant as quiescent WT cells to heat shock (HS). Dilution assays are shown (starting at DO600nm = 5 and diluted 1/5 each time). Left: growth control of exponential cells or 24 h LD cells. Middle: sensitivity to HS of WT exponential cells, 24 h LD cells or 24 h HD cells. Right: sensitivity to HS of WT or sir3∆ LD cells. d Stationary WT, sir3∆, sir3-A2Q cells are resistant to HS like WT cells. Dilution assays are shown (starting at DO600nm = 1 and diluted 1/5 each time). Left: growth control. Middle: 30 min 52 °C HS. Left: 1 h 52 °C HS. e Stationary WT and sir3∆ cells that spent 14 days in water after glucose exhaustion show the same extent of thermotolerance to a 1 h 52 °C HS. Dilution assays are shown (starting at DO600nm = 1 and diluted 1/5 each time). (PDF 12384 kb)

Effect of smoothing on replication curve shape. The chromosome G replication curve is represented after smoothing with different span values, from 0.0001 (almost no smoothing) to 0.075 (extensive smoothing). With low levels of smoothing, several peaks or shoulders did not correspond to ARS positions and were clearly artifacts (0.015 span value). This effect was particularly pronounced for small chromosomes. With extensive smoothing (0.075 span value), close replication peaks tended to merge into one single peak, like, for example, the large left-most double (or triple) replication peak. Comparison of different smoothing levels with ARS positions (red dots) led us to use a 0.04 span value for all chromosomes. (PDF 1101 kb)

Replication timing of histone genes. (DOC 49 kb)

rDNA locus replication. Top: One tandem repeat unit of the S. cerevisiae rDNA locus on chromosome 12 is represented. Coordinates are shown in boxes, according to the Saccharomyces Genome Database (release 19 November 2012). The ARS present in each repeat unit is indicated as a blue diamond. Bottom: Same representation for the C. glabrata rDNA locus. Coordinates are shown according to GĂŠnolevures database (release 10 September 2008). The two ARSs captured are shown by blue diamonds. (PDF 65 kb)

Additional experiment performed to confirm replication origins. The time course shown in this experiment was performed as described in “Methods” for the first experiment (Additional file 1). Top: FACS analyses and DNA content are shown. Time points T0 to T6 were sequenced on a MiSeq (Illumina) sequencer and 2.3–4.1 millions reads were obtained (175 bp single reads) for each time point. Bottom: Sequence coverage for time points T2, T3, T4, and T6, for each chromosome (smoothing span: 0.05–0.09, depending on chromosome size). Sequence mapping of two time points (T1 and T5) exhibited too many gaps to allow us to calculate non-linear regressions for each of the 12 millions nucleotides of the C. glabrata genome, in order to determine T50 values. However, using T6 and T0 coverage, which correspond respectively to S and G1 phases of the cell cycle, we were able to determine replication peak positions. A comparison of peaks detected in experiment #1 (Fig. 2) and experiment #2 (the present figure) was made in the bottom right table. Out of 83 bona fide origins, 73 were found in experiment #2 (88 %), and 241 replication origins were detected (instead of 253 in experiment #1). The average distance between bona fide replication peaks found in both experiments was 6.9 ± 2.4 kb (95 % confidence interval), consistent with what was deduced from comparisons between ARS positions and replication peaks (Fig. 5a). (PDF 1224 kb)

Telomere hyperclustering is not due to slow growth. a Representative fluorescent image of Rap1-GFP tagged strain grown either at 30 °C or 25 °C in exponential phase (top) and then starved for 16 h in water before imaging (bottom). b Calcofluor staining of LD and HD fractions of a post DS culture after gradient separation. c Heat shock (HS) assay on the LD and HD fractions used in b. (PDF 11591 kb)

Characterization of the SP silent chromatin hypercluster. a Western blot against Rap1 on crude extracts from exponential, respiratory, or stationary cultures of a WT strain (yAT1684). H2A antibody was used for the loading control. b Representative fluorescent images of wild-type (WT) strains tagged with Rap1-GFP “yAT 1684”, GFP-Sir2 “yAT405”, Sir3-GFP “yAT779” and GFP-Sir4 “yAT431” strains. Overnight liquid cultures were diluted to 0.2 OD600nm/ml and images were acquired after 5 h (1 OD600nm/ml, fermentation phase) and 7 days (40 OD600nm/ml, stationary phase). c Representative fluorescent image of a Rap1-GFP Sir3-mCherry-tagged strain “yAT194” from stationary phase cultures. We note that Sir3 associates with both telomeres and the rDNA in stationary phase cells. d Representative fluorescent images of Rap1-GFP in stationary cultures of WT “yAT1684” and sir4∆ “yAT2092” strains. e Representative fluorescent images of the nucleolar protein Sik1 tagged with mCherry during fermentation, respiration, and stationary phase (“yAT340”). f Representative fluorescent image of Rap1-GFP Dad2-mRFP (Duo1 And Dam1 interacting, an essential component of the microtubule–kinetochore interface) tagged stationary phase cells (“yAT2279”). g Representative fluorescent image of Sir3-mCherry Cse4-GFP-tagged strain “yAT2280” from stationary phase. Scale bar is 1 μm. (PDF 1343 kb)

Plasmid replication efficacy of DNA regions surrounding identified replication origins. (DOC 40 kb)

This thesis deals with several aspects concerning genetic regulation on theoretical and experimental points of view. A first work in the field of modelisation of biological systems, proposes to find simple regulatory networks which have the ability to optimise their response when they are submitted to a periodic stimulation. The network called ``Incoherent Feed Forward Loop'' appeared to have the interesting property to let pass trains of pulses which have a particular temporal pattern. Some extensions of this motif (``Diamond'', ``Double Diamond''...) are proposed to also have interesting properties to process more complex time-dependent signals. Next, a work of review and observation concerning DNA looping is presented. After pointing out some common points in systems of DNA looping already known, we have examined the data bases to see if other regulatory regions have the same characteristics. We propose a list of several genes which might be good candidates to be regulated thanks to a DNA looping mechanism. An experimental work of molecular biology is then presented. It tackles a hypothesis found in several works from bioinformatics and statistical physics. The question is the following: can transcription factors bind to binding sites belonging not to a unique regulatory region of one gene (like in the lac operon) but to binding sites belonging to different genes? We tested this hypothesis on the Pel regulatory system which rules the entry into virulence of the enterobacteria Erwinia chrysanthemi. Experiments reproduced and carried out in various physiological conditions lead to the conclusion that the deleted regulatory regions do not seem to act on other genes by a DNA looping mechanism. The last chapter proposes an experimental method to search for such interactions in a broader way inside a bacterial genome. The method is based on molecular biology techniques commonly used, like the random mutagenesis or alpha-complementation. It uses the enhancer system of the promoter of glnA of Escherichia coli and aims at finding significant 3D interactions between DNA segments in a prokaryote genome., Cette thèse s'intéresse à plusieurs aspects concernant la régulation génétique sur les plans théorique et expérimental. Un premier travail théorique qui s'inscrit dans le cadre de la modélisation des systèmes biologiques se propose de trouver des réseaux de régulation simples qui puissent répondre de manière optimale lorsqu'ils sont soumis à une stimulation périodique. Le réseau appelé ``Incoherent Feed Forward Loop'' s'est avéré présenter la propriété intéressante de laisser passer des trains de pulses au profil temporel particulier. Des extensions de ce motif (``Diamond'', ``Double Diamond''...) ont été suggérées pouvant également présenter des propriétés intéressantes pour traiter des signaux plus complexes. Un travail de revue et d'observations concernant les boucles d'ADN est ensuite présenté. Après avoir observé des points communs dans les systèmes de boucle d'ADN déjà connus, nous avons interrogé les bases de données pour savoir quelles étaient les régions de régulation présentant les mêmes caractéristiques. Nous proposons une liste de plusieurs opérons qui mériteraient une démarche expérimentale pour la mise en évidence d'une boucle d'ADN. Un travail expérimental de biologie moléculaire est ensuite présenté. Il s'est attaqué à tester une hypothèse présente dans plusieurs travaux de bioinformatique ou de physique statistique. La question testée est la suivante: est-ce que des facteurs de transcription pourraient se lier à des sites de liaison appartenant non pas à la même région de régulation d'un gène (comme dans le cas de l'opéron lac) mais à des sites de liaison appartenant à des gènes différents? Nous avons testé cette hypothèse sur le système de régulation d'entrée en virulence de la bactérie Erwinia chrysanthemi. Les expériences reproduites et réalisées dans plusieurs conditions physiologiques différentes ont abouti à la conclusion que les régions de régulation supprimées ne semblent pas agir sur d'autres gènes par le mécanisme de boucle d'ADN. Le dernier chapitre propose une méthode expérimentale pour rechercher de telles interactions de façon large dans un génome de bactérie. La méthode est basée sur des techniques de biologie moléculaire couramment utilisées comme la mutagénèse aléatoire ou l'alpha-complémentation. Elle utilise le système enhancer du promoteur de glnA d'Escherichia coli et vise à trouver des interactions 3D significatives entre segments d'ADN au sein d'un génome de procaryote.

International audience; A computational challenge raised by chromosome conformation capture (3C) experiments is to reconstruct spatial distances and three-dimensional genome structures from observed contacts between genomic loci. We propose a two-step algorithm, ShRec3D, and assess its accuracy using both in silico data and human genome-wide 3C (Hi-C) data. This algorithm avoids convergence issues, accommodates sparse and noisy contact maps, and is orders of magnitude faster than existing methods.