Thèse Rôle de la Topologie Locale du Chromosome dans la Co-Expression de Gènes Voisins chez les Bactéries H/F

Doctorat.Gouv.Fr

Établissement : INSA Lyon École doctorale : E2M2 - Evolution Ecosystèmes Microbiologie Modélisation Laboratoire de recherche : MAP - MICROBIOLOGIE, ADAPTATION ET PATHOGENIE Direction de la thèse : Sam MEYER ORCID 0000000172066226 Début de la thèse : 2026-10-01 Date limite de candidature : 2026-06-09T23:59:59 Cette thèse porte sur l'influence de l'organisation physique du chromosome bactérien sur l'expression des gènes. Dans les génomes bactériens, des gènes voisins présentent souvent des niveaux d'expression corrélés, même sans régulation transcriptionnelle classique. Certains regroupements de gènes, appelés « segments de synténie » ou îlots de pathogénicité, restent conservés au cours de l'évolution, suggérant un rôle important de l'environnement chromosomique local.
Le projet vise à comprendre quantitativement comment la topologie locale de l'ADN - notamment le surenroulement et l'organisation en domaines topologiques - influence l'activité des gènes voisins. Pour cela, l'étudiant combinera des données transcriptomiques (RNA-Seq) et des données de structure 3D du chromosome obtenues par microC, une version très haute résolution de la méthode HiC. Les travaux seront menés sur deux bactéries modèles : Escherichia coli et Dickeya dadantii, une bactérie phytopathogène.
Dans E. coli, l'étude exploitera des données inédites où une boucle artificielle de 7 kb a été créée dans le chromosome afin d'analyser son impact global sur l'expression des gènes. Dans D. dadantii, la thèse débutera par l'adaptation expérimentale de la méthode microC afin d'étudier le rôle de protéines architecturales comme H-NS et de la topologie de l'ADN dans l'expression des facteurs de virulence.
La majeure partie du doctorat sera consacrée à l'analyse computationnelle et à la modélisation dynamique des interactions entre transcription, surenroulement de l'ADN et protéines associées au chromosome. L'objectif est de construire un modèle prédictif capable d'expliquer comment l'environnement génomique local module l'expression des gènes. Ce travail, à l'interface entre biologie moléculaire, physique et bioinformatique, présente des applications importantes en microbiologie et en biologie synthétique. Bacterial genomes are very dense, and neighboring genes often exhibit correlations of expression that cannot be explained by classical transcriptional regulators [Junier, 2016]. Some of them, called synteny segments, show unusual conservation of this proximity along evolution even when they seem functionally unrelated. A particular case is pathogenicity islands, i.e., clusters of virulence genes in pathogenic bacteria.
The role of local chromatin structure, and in particular of chromosome topology, in coordinating gene expression within topological domains of 10-100 kb, has been highlighted by several studies [Lioy, 2018; Gavrilov, 2025; Boulas, 2023]. However, this observation remains qualitative, and does not allow any quantitative prediction of how a gene's genomic environment affects its expression and vice versa. Such a quantitative model of neighboring gene interaction is of major importance, both as a fundamental question and for applications. In synthetic biology, it is necessary to ensure a controlled expression when a gene of interest is inserted into the chromosome of a bacterium [Yeung, 2017]. In pathogenic bacteria, it is recognized as a key factor in the expression of virulence functions [Dorman, 2016; Martis, 2019]. For example, it was recently shown that positive DNA supercoiling resulting from the expression of nearby genes can disrupt the binding of H-NS, a major architectural protein affecting the topology, in a Salmonella enterica pathogenicity island, and thus trigger its expression [Figueroa-Bossi, 2024]. This project aims at building a quantitative and dynamical modeling of how neighboring genes interact via local topology. It will be obtained by a combination of RNA-Seq data and a new highly resolved method to monitor the local topology of the chromosome called microC, in two related enterobacteria, Escherichia coli and Dickeya dadantii. MicroC is a recent update of the HiC conformation capture methodology [Lioy, 2018] that allows a dramatic improvement in spatial resolution, from 1-5 kb to ~10 nt, thus giving access for the first time to most features relevant for transcription and its regulation [Gavrilov, 2025]. While quantitative models of transcription-topology coupling were proposed, including at the host laboratory [El Houdaigui, 2019; Boulas, 2023; Goldberg, 2025], these new data will give an essential input to include the binding and unbinding of topologically-related proteins (nucleoid-associated proteins, especially H-NS) in this modeling, which is believed to be a major factor in vivo. In E. coli, the student will analyze already obtained but unpublished data (RNA-Seq and microC) obtained by a collaboration of the host team with the Jie Xiao lab at Johns Hopkins university, where a 7 kb-loop was artificially induced in the chromosome (using the lambdaCI protein). This topological change affected the expression of genes within and near the loop in a complex way, depending on their orientation [Yehya, 2025]. Recently, RNA-Seq and microC data were obtained in these strains, and show that the loop has profound effects on the expression of genes along the whole chromosome, which will be combined and analyzed here.
Dickeya dadantii, a plant pathogen studied in the host laboratory, is a model organism for the role of local topology in virulence functions, dependent on DNA supercoiling and H-NS, as shown experimentally in the host team [Ouafa, 2012], but there is a comparable lack of an encompassing model. In recent months, the team has accumulated a wealth of RNA-Seq and H-NS ChIP-Seq data that provide a detailed picture of these processes. The student will start the PhD by adapting the microC protocol to that bacterium (WT and hns mutant, in presence or absence of topoisomerase inhibitors novobiocin and seconeolitsine), under the guidance of the host team, experts of Dickeya, and Virginia Lioy, expert in HiC who has also adapted the microC methodology in E. coli and S. enterica (estimated time: 6 months).
The rest of the PhD will consist in computational analyses and modeling, starting by combining the different sources of data to compare the mechanisms identified in the two species, and compare the effect of the artificial loop by lambdaCI versus natural topological constraints imposed by H-NS. This analysis will be combined with dynamical modeling of the process. Starting from a simulation code recently developed to simulate this process with a fixed topological barrier [El Houdaigui, 2019; Goldberg, 2025], the model will be updated to include the supercoiling-dependent binding and unbinding of H-NS and lambdaCI.